MASARYKOVA UNIVERZITA Proceedings of the 10th Summer School of Applied Informatics Sborník příspěvků 10. letní škola aplikované informatiky Bedřichov 6.–8. 9. 2013 Editoři: Editors: Techničtí editoři: Technical editors: Jiří Hřebíček Jan Ministr Tomáš Pitner Jiří Kalina Radka Novotná Odborní garanti letní školy: Prof. RNDr. Jiří Hřebíček, CSc., MU, IBA Expert guarantos: Doc. RNDr. Tomáš Pitner, Ph.D., MU, FI Programový výbor: Editorial and programme board: Doc. RNDr. Ladislav Dušek, Dr. – Masarykova univerzita, Brno, CZ Doc. Ing. Josef Fiala, CSc. – Vysoká škola podnikání, Karviná, CZ Doc. Ing. Alena Kocmanová, CSc. - Vysoké učení technické v Brně, CZ RNDr. Jaroslav Ráček, Ph.D. – IBA CZ, s.r.o., CZ Prof. Andrea Rizzoli – University of Applied Sciences and Arts of Southern Switzerland, Lugano, CH Dr. Gerald Schimak - Austrian Institute of Technology, Vienna, AT Doc. Ing. Jan Žizka, CSc. – Mendelova univerzita v Brně, CZ © 2013 Nakladatelství Littera ISBN 978-80-210-6058-6 Slovo úvodem 10. letní škola aplikované informatiky navázala na předchozí letní školy aplikované (environmentální) informatiky v Bedřichově, které se zde konají od roku 2002 s výjimkou roku 2004, kdy se letní škola konala v Šubířově a roku 2005, kdy byla hlavní akcí Masarykovy univerzity (MU) v oblasti environmentální informatiky na 19. mezinárodní konferenci Informatics for Environmental Protection - EnviroInfo 2005 s nosným tématem Networking environmental information a kterou hostila Masarykova univerzita ve dnech 7. až 9. září 2005 v brněnském hotelu Voroněž. V letech 2006 až 2012 se letní školy konaly opět v Bedřichově v penzionu U Alexů. V letošním roce 10. letní škola aplikované informatiky se konala ve dnech 6. až 8. září 2013 v Bedřichově opět v Bedřichově v penzionu Artis (dříve U Alexů). Všemi přednáškami 10. letní školy prolínala skutečnost, že aplikace moderních informačních a komunikačních technologií pro životní prostředí (potažmo environmentální informatiky) jak v České republice (ČR), tak i mezinárodně v Evropské unii (EU) a ve světě se zaměřuje na podporu eEnvironmentu, Jednotného informačního prostoru pro životního prostředí (SISE – Single Information Space in Environment for Europe) a Sdíleného informačního systému pro životní prostředí (SEIS – Shared Environmental Information System), které podporují naplňování nové politiky v budování informační společnosti EU a ČR, které přinesla „Digital Agenda for Europe“ v rámci vize nové Evropské komise „eEUROPE 2020“. Jde zejména o přeshraniční informační služby v rámci eEnvironmentu. Jedná se o sdílení monitorovaných a zpracovávaných dat a informací o atmosféře, povrchových i podzemních vodách, odpadech, půdě, biodiverzitě, atd. pomocí Globálního monitorovacího systému životního prostředí a bezpečnosti (GMES – Global Monitoring for Environment and Security). Tyto informační služby umožňují efektivnější a přesnější sledování aktuálního stavu životního prostředí a udržitelného rozvoje v Evropě, dále pak jeho modelování a simulaci jeho dalšího vývoje. Za hlavní přínos 10. letní školy považujeme skutečnost, že na letošním ročníku setkali a jsou v příspěvcích zastoupeni nejen doktorandi a učitelé z Masarykovy univerzity (Institut biostatistiky a analýz, Přírodovědecká fakulta a Fakulta informatiky), ale i z Mendlovy univerzity (Provozně ekonomická fakulta), Vysokého učení technického (Podnikatelské fakulta) v Brně a Vysoké školy báňské – Technické univerzity (Ekonomická fakulta) v Ostravě. Jejich příspěvky ve sborníku přispívají k tomu, že se letní škola stala širokým interdisciplinárním odborným fórem v rámci České republiky. Dále je důležité, že několik příspěvků, jejichž spoluautoři jsou ze zahraničí, je publikováno v anglickém jazyce, který podtrhuje mezinárodní význam sborníku. Těžiště projednávaných otázek na letní škole bylo především v detailní diskusi věnované řešení projektů vývoje informačního systému GENASIS, ale zahrnulo i další problematiku, která se týkala oblasti „eEnvironment“, „eGovernment“, „eParticipation“ a prezentaci nově řešeného projektu podporovaného Grantovou agenturou České republiky s názvem „Construction of Methods for Multifactor Assessment of Company Complex Performance in Selected Sectors”, vedený pod registračním číslem P403/11/2085. V Brně dne 31. listopadu 2013 Jiří Hřebíček Jan Ministr Tomáš Pitner Editoři Obsah Cross Platform Mobile Development Jakub Dubrovský 1 Business Intelligence in Environmental Reporting Powered by XBRL M. Hodinka, M. Štencl, J. Hřebíček, O. Trenz 15 Sustainability indicators for measuring SME performance and ICT tools Edward Kasem, Jiří Hřebíček 26 Curriculum planning and construction: How to combine outcome-based approach with process-oriented paradigm Martin Komenda, Lucie Pekárková 39 Information Technology Tools for Current Corporate Performance Assessment and Reporting Ondřej Popelka, Michal Hodinka, Jiří Hřebíček, Oldřich Trenz 47 Generování japonsko-českého slovníku Jonáš Ševčík 69 Availability and scalability of web services hosted in Windows Azure David Gešvindr 73 Support for Partner Relationship Management Internet Data Sources in Small Firms Jan Ministr 84 Cross Platform Mobile Development Jakub Dubrovský Masarykova univerzita - Fakulta informatiky Botanická 68a, 602 00 Brno jkdubr@me.com Abstract Mobile devices are fast becoming ubiquitous in today’s society. New devices are constantly being released with unique combinations of hardware and software, or platforms. In order to support the ever-increasing number of platforms, developers must embrace some method of crossplatform development in which the design and implementation of applications for different platforms may be streamlined. To solve that, cross platform tools have been investigated and was found to solve device functionality features, platform and device coverage problems. It was also noted that cross platform development was found to be cheaper and takes less time. Abstract Mobilní zařízení se rychle stávají všudypřítomné v dnešní společnosti. Nová zařízení jsou vydávány s neustálou unikátní kombinací hardwaru a softwaru nebo platformy. S cílem podporovat stále rostoucí počet platforem, vývojáři musí přijmout nějakou metodu crossplatform vývoje, v němž může zefektivnit návrh a implementaci aplikací pro různé platformy. Vyřešit, multiplatformní nástroje byly prošetřeny a zjistilo se, že řešení funkce zařízení nabízí zařízení a platformu pokrytí problémy. Bylo také zjištěno, že multiplatformní vývoj je levnější a trvá kratší dobu. Keywords Mobile development, Android, iOS, Windows Phone, Phonegap, Cordova Klíčová slova Vývoj mobilních, Android, iOS, Windows Phone, Phonegap, Cordova 1 Introduction Today mobile applications are in a lot of sense quite—with the right and cleverly designed applications—capable of pulling above their weight to compete with traditional stationary computers. Applications can be developed using a wide variety of tools, methodology and using the traditional applications development life cycle. The number of tools that promise a trouble free, squeaky clean UI, lightning speed development period, and maximum interoperable software product can only be described as staggering. The mobile applications development scene has split into at least two categories—native and cross platform development. The most straightforward method of developing an application will be to target native applications with native thinking, native target and deliberately native technology. Besides it runs faster, most of its ailments are probably fully groped through, well documented idiosyncrasies and often a relatively reliable but expensive professional support. One of the major roadblocks to native development is the number and variety of platforms and mobile devices in the market—making mobile applications development end up with fragmented development arrangements and costly software development strategies [1]. The potential benefits of writing-once-running-everywhere comes from the cost reduction—in having only one code to write and maintain, and the time reduction— being able to write one code and target multiple devices and platforms, making researching cross-platform mobile applications development worth one’s effort and findings The main objectives of research questions are: Is native development faster than cross platform development? What is the benefit of using cross platform SDKs to develop cross platform? Evaluate cross platform tools using time, technology, maturity, and cost aspects of mobile apps development. What tool best supports cross-platform development? 2 Mobile platforms 2.1 Android mobile operating system Android is a Linux-based operating system designed primarily for touchscreen mobile devices such as smartphones and tablet computers. Initially developed by Android, Inc., which Google backed financially and later bought in 2005, Android was unveiled in 2007 along with the founding of the Open Handset Alliance: a consortium of hardware, software, and telecommunication companies devoted to advancing open standards for mobile devices. The first Android-powered phone was sold in October 2008 [7]. 2.2 Apple iOS mobile operating system The Apple iPhone hit the market in January 2007 and brought with it the iOS operating system—an operating system filled with goodies that seemed to tickle everyone’s fancy. iOS is a heavily guarded proprietary operating system whose fate is tied to a romanticized device in the name of iPhone and other iFamily devices such as iPod touch, and iPad devices. Before the advent of Android, the iPhone proved to be a formidable weapon in the war of smartphone dominance and successfully positioned iOS itself in the top rank in just a short time since its debut, now just trailing behind Symbian and in 2007, it was considered as the invention of the year in five categories among which were being beautiful and defining the future [2]. iOS provides services that any operating system provides for iFamily devices. In addition to its pioneering technology and innovative design, iOS was considered to be a status symbol making it a market leader in a short period of time. Consequently, other operating systems in the market had to make way for it and die. One of the obstacles to developing applications on iOS is that it necessarily required a license for developers and that the system is closed and that the apps are Apple approved products on the iGlue online store of iTunes app store, and not to mention the that fact that the devices come at a premium. 2.3 Windows Phone 8 Windows Phone 8 is the second generation of the Windows Phone mobile operating system from Microsoft. It was released on October 29, 2012, and like its predecessor, it features the interface known as Metro [9]. Windows Phone 8 replaces its CE-based architecture used on Windows Phone 7 devices with the Windows NT kernel found on many Windows 8 components. Current Windows Phone 7.x devices cannot run or update to Windows Phone 8 and new applications compiled specifically for Windows Phone 8 are not made available for Windows Phone 7.x devices [9]. 2.4 Mobile platforms with low market penetration Most popular mobile operation system is iOS and Android is operating system with the biggest market penetration. What about other platforms? Another operation system is for example Bada by Samsung, BlackBerry. These platforms have low market penetration and developers have no big reason for developing for these platforms. 2.5 Common native features of a mobile platform There are several native features in a mobile phone. Some of the most common native features supported in a mobile are described below. Accelerometer is a feature of a mobile phone that is used to detect the orientation of a mobile device or how a user holds the phone. It can be used to detect motion in three dimensions. Geolocation is an API that is used to provide or access geographical location of a mobile holder or the mobile device to a computer application. Location information is usually based on Global Positioning System (GPS), IP address, Wi-Fi, Bluetooth, MAC address or GSM/CDMA mobile networks. Barcode is an API used to read barcodes through a mobile device camera and interpret the information encoded in the barcode. The Calendar API gives access to the native calendar functionality and data of a mobile phone platform. A fast lightweight embedded SQLite database module in a mobile platform. It is used to create, access and store data on the mobile phone and accomplish database management tasks. The media feature enables to access multimedia functionalities of a mobile platform such as playing and recording of audio and video. Access to platform players and multimedia file operations. Notification allows raising audible, visual or vibrational notification. Push notification services are often based on information preferences expressed in advance. This is called a publish/subscribe model. A mobile device "subscribes" to various information "channels" provided by a server; whenever new content is available on one of those channels, the server pushes that information out to the client. 3 Mobile development 3.1 Native mobile applications development Native application development involves developing software specifically designed for a specific platform, hardware, processor and its set of instructions. The advantages of native applications development are predictable performance, streamlined support, native UI, native API, coherent library updates. 3.2 Cross platform development Cross platform development promises lower cost and time of development since it produces one code base to maintain and write targeting multiple devices and platforms. The figure below shows the trend for native to cross platform development cost and time factors. Figure 1: Native vs. Cross platform development [1] Claims that cross platform development might inhibit future development in the field since it might be focused only on making it easy for developers than on innovation. Some benefits of cross platform development are: Learning web technologies might be less challenging that learning an all new platform Devices come in all sorts of technology and feature that through cross platform development it might be possible to develop an application that works well on each one of them with less effort Cross platform mobile apps written using one SDK, operating System and development environment requiring less financial investment Cross platform apps try best to emulate natively written applications Cross platform development might help reduce cost in human resource, time resource, and development period Table 1: App development comparison 4 Cross platform tools 4.1 Appcelerator Titanium Appcelerator Titanium is a cross platform mobile application development tool. It claims to use open web standards, architecture and follows open source initiatives. Titanium Mobile SDK comes with its own IDE. Titanium Studio is an eclipse like IDE which is used to write and test applications together with and Android Emulator. Appcelerator titanium supports database, media, geolocation, contacts, notification and many other native features of a smartphone. Titanium enables web developers to create native mobile, desktop, and tablet applications using open web technologies such as JavaScript, HTML and CSS. One uses JavaScript APIs for UI (native), JavaScript for scripting 4.2 PhoneGap PhoneGap is a cross platform mobile applications development tool. It uses HTML5, JavaScript and CSS3 web technologies to develop cross platform mobile applications that exploit native features of a mobile device [8]. PhoneGap supports the largest number of platforms of all other tools—iOS, Android, BlackBerry, webOS, WP7, Symbian and Bada. PhoneGap developed apps are hybrid apps that are neither pure web apps nor native apps. PhoneGap is an open source mobile apps development tool. With PhoneGap, one can develop an abstraction based cross platform applications using web technologies and wrap the code in native accessing system architecture of an application [8]. PhoneGap uses jQuery JavaScript library in its development framework and made it easier to build jQuery Base mobile apps to native feature accessing applications. It supports accelerometer, camera, compass, contacts, file, geolocation, media, network, notification (alert, sound, vibration) and storage [8]. 4.3 Xamarin Xamarin is a commercial cross platform mobile applications development tool. Xamarin enables to develop applications for iOS and Android using .NET framework and C# programming language. Application development using Xamarin comes in two forms as MonoTouch for iOS and Mono for Android [3]. Mobile applications development using Xamarin is claimed to bring the benefit of sharing codes between platforms, using existing .NET expertise, easy access to native APIs, the niceties of rich IDE—Visual Studio in case of Android—and developing applications using strongly typed programming along with garbage collecting feature of .NET framework. Xamarin IDE comes as MonoDevelop for android and as a Visual Studio plug-in component for windows and Mac OSX and as monotouch for Mac OSX only [3]. 4.4 Rhombile Rhomobile is a Motorola Solutions owned company that brought a cross platform mobile applications development tool that relies on Model-View-Controller(MVC) system architecture of programming applications using HTML, CSS and JavaScript and Ruby. Rhomobile supports iOS, RIM, Windows Mobile and Android [4]. 4.5 MoSync MoSync is an open source cross platform mobile applications development tool. It enables one to develop native like cross platform mobile applications using C/C++, HTML5 and JavaScript [5]. 4.6 Corona Corona is a cross platform mobile applications development tool that is used to author games and apps. Mobile applications development in Corona are written using the programming language Lua. The Lua code is compiled into an intermediate bytecode for a native runtime abstraction 5 Conclusion Cross platform mobile applications come in so many forms. Each tool might approach cross platform mobile applications development in so many ways. Some come as runtime implementations or interpreter abstraction layers and when it comes to licensing some are completely free and open source while some are proprietary and commercial. The level of native feature support of a mobile device varies widely that much. As seen in the previous sections, the level native support some tools provide is wider and some of the other ones also provide a limited support but might also provide a special feature that a particular developer might be keen to utilize. Cross platform development is an actively developing young technology. It could be sometime before the dust settles and a tools comes out as a winner. Considering each tool based on its merit for a mobile application need and target audience seems the only sensible option available right now. 6 References [1] AllianceTek. Comparison chart for mobile applications development, 2011 http://www.alliancetek.com/downloads/article/comparison-chart-for-mobileapp.pdf, April 2012. [2] Feida Lin and Weiguo Ye. Operating system battle in the ecosystem of smart- phone industry. In Information Engineering and Electronic Commerce, 2009. IEEC ’09. International Symposium on, pages 617 –621, may 2009. [3] Xamarin. Xamarin documentation http://www.xamarin.com, August 2013. [4] Rhomobile. The universal mobile application http://www.rhomobile.com/products/rhoconnect/, August 2013. integration server [5] MoSync. Mosync home http://www.mosync.com/, August 2013. [6] Badadevelopers.Architectureofbadahttp://developer.bada.com/library/help, August 2013. [7] Android. Android https://developer.android.com/index.html, August 2013. [8] Phonegap. Phonegap for developers http://phonegap.com/developer/, August 2013. [9] Windows Phone. Dev Center for developers http://developer.windowsphone.com/en-us, August 2013. Business Intelligence in Environmental Reporting Powered by XBRL M. Hodinka, M. Štencl, J. Hřebíček, O. Trenz Mendel University in Brno, Department of Informatics, Faculty of Business and Economics, michal.hodinka@mendelu.cz, michael.stencl@mendelu.cz, jiri.hrebicek@mendelu.cz Abstract Today, companies are handling increasing amounts of transactional data. This phenomenon commonly named as "Big Data", has transformed from a vague description of massive corporate data to a household term that refers to not just volume but the variety, veracity and velocity of change. Commonly used approach leads to usage of Business Intelligence (BI) technologies used not only to environmental reporting purposes, but also used for a data discovery discipline. The critical issue in general data processing tasks is to get the right information quickly, near to real time, targeted, and effectively. This article aims on several critical points of the whole concept of BI environmental reporting powered by XBRL. First, and most important, is the usage of structured data delivered via XBRL. The main profit on usage of XBRL is the optimization of the ETL process and its combination commonly used best practices on data warehouse models. The whole BI workflow could be moved further by additional data quality health checks, extended mathematical and logical data test, basics of data discovery and drill-down techniques. First part of the article review the state of the art on the XBRL level and also review current trends in environmental reporting. We also analyse the basics of Business Intelligence regarding to the application domain on environmental reporting. The methodology reflects today’s technical standards of XBRL accordingly to the application via ETL process. In results we describe concept for standardized data warehouse model for the environmental reporting based on the specific XBRL taxonomy and known dimensions. In discussion we explain our next approach and all the pros and cons of the selected approach. Abstrakt Společnosti mají dnes větší množství transakčních údajů k zpracování. Tento jev běžně pojmenovaný jako "Big Data", se přeměnil z vágní popis masivní podnikových dat na domácností termín, který odkazuje na nejen objem, ale odrůdu, pravdivost a rychlost změn. Běžně používaný přístup vede k využití Business Intelligence (BI), technologie nejen pro environmentální účely vykazování, ale také pro disciplínu zjišťování údajů. Kritickou otázkou obecně zpracování dat úkolů je dostat správné informace rychle, v blízkosti reálném čase, cíleně a efektivně. Tento článek se zaměřuje na několik kritických bodů celého konceptu BI environmentálního reportingu poháněné XBRL. První a nejdůležitější je využití strukturovaných dat prostřednictvím XBRL. Hlavní zisk z použití XBRL je optimalizace procesu ETL a jeho kombinace běžně používají osvědčené postupy v datový sklad modelů. Celý pracovní postup BI mohl být dále přesunutý, další údaje kvality kontroly, test rozšířené matematických a logických dat, základy datových objevu a podrobnostem techniky. První část tohoto článku zkontroluje stav techniky na úrovni XBRL a rovněž přezkoumává současné trendy v environmentálního reportingu. Také jsme analyzovali základy analytické nástroje týkající se domény aplikace na environmentálního reportingu. Metodika odráží dnešní technické standardy XBRL odpovídajícím způsobem k aplikaci prostřednictvím procesu ETL. Výsledky popisujeme jako koncept pro standardizované datový sklad model pro ekologické vykazování založené na určité taxonomie XBRL a známých dimenzí. V diskusi si vysvětlíme, náš další přístup a všechny výhody a nevýhody vybraných přístupu. Keywords Corporate reporting, Environmental reporting, XBRL, Business Intelligence, Performance evaluation, Key performance indicators Klíčová slova Podniková vykazování, environmentální reporting, XBRL, Business Intelligence, hodnocení výkonnosti, klíčové ukazatele výkonu 1 Introduction Big data as a term is one of the most mentioned topic in past years. As a major factor standing behind the massiveness of a big data processing in the SMB is the easy access to techniques and tools that help to any user generate information from any data. The current product philosophy of technological giants as Google or Apple already changed the IT and software world. The “user friendliness” or “user experience” of current products allows using a complex data analytics tools in very easy manner without a deep analytic and reporting knowledge and skills. So, the world of analytics had already changed. Developers are now facing new goals or old known “hard-to-handle” tasks like data quality, real-time analytics of streaming data, different forms and uncertainty of data, reducing the complexity of data transformations and others. Commonly we can speak about the Business Intelligence (BI). BI technologies are widely used not only to environmental reporting purposes, but those technologies are part of the data discovery discipline. Traditional understanding defines Business intelligence, or BI, as an umbrella term that refers to a variety of software applications used to analyse an organization’s raw data. BI as a discipline is made up of several related activities, including data mining, online analytical processing, querying and reporting and has been adopted mostly by the enterprise companies, not the SMB segment. Today, also the SMB companies are handling increasing amounts of transactional data and they have an easy access to even bigger scale of data from e.g. social networks and other “new” media resources. This phenomenon could also be classified into "Big Data". Big data now changed its definition from a vague description of massive corporate data to a household term that refers to not just volume but the velocity, veracity or variety of data. The critical issue in data processing and data analysis tasks is to get the right information quickly, near to real time, targeted, and effectively. Speaking about traditional approach (data marts), and new in-memory analytics with main motivation to explain the structured/unstructured data in order to real-time analytics and data efficiency, aka data monetization. The purpose of this paper is to enhance understanding of the current approach to BI along with the extension based on the extensible business reporting language (XBRL) tools ready to transform existing data into an informational source of knowledge. It begins by linking cutting-edge scientific research on the XBRL level with review of current trends in environmental reporting. The analysis includes basics of Business Intelligence regarding to the environmental reporting. It serves the purpose to show how complexity can change by applying an XBRL data model. 2 Methods and Resources The methodology reflects today’s technical standards of XBRL accordingly to the application via ETL process. In results we describe concept for standardized data warehouse model for the environmental reporting based on the specific XBRL taxonomy and known dimensions. In discussion we explain our next approach and all the pros and cons of the selected approach. Any reporting service must be based on a set of predefined industry specific KPIs. The Key Performance Indicators (KPIs) for the measurement of economic performance in relation to the sustainability and ESG indicators. The economic performance indicators provide quantitative forms of feedback, which reflect the results in the framework of corporate strategy. The approach is not different when we control environmental, social and governance issues. The non-financial KPIs that an organization develops, manages and ultimately reports – whether internally or externally – will depend on its strategic priorities, and will reflect the unique nature of the organization. What is most important is to recognize what is measured, what is controlled, and it is important that the measures create value for the company and its stakeholders. The proposed KPIs can help organizations to plan and manage their economic priorities, in particular, when the economic indicators are focused on the core business strategy, by means of operational plans, which include performance targets. We want to demonstrate, how XBRL can be potentially applied in different areas beyond the original design objectives of the standard. Many organizations have focused on employing XBRL GR in primarily transaction-oriented focus. How to link integrated system with external reporting using XBRL shows us Klement (2007). Standard reporting process takes XBRL GL instance and import into central data warehouse. This instance is imported into a relational database, which serves basis for incremental updates to data warehouses and then to the OLAP cubes respectively. Currently Klement (2007) think there isn't any XBRL-based standard tools for this purpose. To avoid large quantities of transactions exist highly optimized bulk load toolkits for data import into relational databases but XBRL-based implementation will not replace performance optimized bulk load toolkits in the short term. The severest problem at the so-called drill-down or drill-around is likely to be data access over system boundaries. XBRL reports also have the capacity to incorporate benchmarking and drilldown capabilities to access highly detailed level information on a ”need to know” basis. A significant advancement promises is a central XBRL repository for storing reporting facts and data mapping. An important prerequisite to improve drill-down functions for aggregation is no inverse drill-down function back to the original facts and available of mapping descriptions, which support efficient analysis queries. For that purpose Klement (2007) did not find any suitable XBRL standards. In this example there exist three transitions between non-XBRL and XBRL formats: 1. ERP data export to XBRL GL 2. XBRL GL import into relational database (a typical ETL application) 3. OLAP cube data export to an XBRL FR instance BI provides a broad set of algorithms to explore the structure and meaning of data. All the data scrubbing and pre-processing (extract, transform and load: ETL) has to do with mapping of metadata and can be neglected when leverage clean and meaningful XBRL data. Chamoni (2007) ask the question: why not use semantic layer and taxonomy of XBRL to go beyond reporting and do more in-depth analysis? Real-time control of business processes is currently hyped within the data warehousing industry. As every business process should possibly be traced in the accounts of a company a constant flow of financial data in XBRLformat into BI-system will be necessary for continuous control of operations, for early fraud detection and BI as a source of compliance systems. The future enterprise will be based on the intelligent real-time technologies. Chamoni (2007) point out what future research must be done to develop analytical applications with a high degree of intelligence and very low reaction time based on XBRL and BI. The contribution of XBRL in BI is semantically enriched data transport in favor of generic base systems and may be used to build presentation and access systems. Fig. 1: Shell of BI-tools (Gluchowski, Kemper 2006) Modern BI solutions do not longer update data in periodic ETL process, but use event-based triggers to feed the analytical applications with streaming of data. Short response time and synchronous signal processing opens the field for new control applications. The significance of these reactive BI solutions is significant and Chamoni (2007) shown the impact in research of framework for BI and XBRL. It leads to an issue of having different multidimensional modeling approaches: one for data warehouses and another for business reports, which are usually transferred between data warehouses. There is a new multidimensional approach based on XBRL Dimensional Taxonomies (XDT). This add on described by Felden (2007) has the potential to perform highly sophisticated multidimensional tasks such as directly facilitating OLAP solutions. Felden (2007) can be continued by different research lines. It can be related to other areas like data quality, database security, temporal issues, query optimization, and translation to logical/physical level methodologies, or just studying modeling problems at conceptual level. Especially the conceptual level offers research potential. In XBRL is data model based on taxonomies expressing metadata and instance documents referring to the taxonomies representing business reports. The primary taxonomies represent business fact data, which are later reported accordingly in instance documents. The domain member taxonomies model the content of the explicit dimensions and the holder properties for the typed dimensions. The template taxonomies amend the multidimensional model connections the primary items with dimensions using hypercubes (HOFFMAN, 2006). In order to model the relationships between various elements (primary items, hypercubes, dimensions and domain members) the following connections (arcroles) (PIECHOCKI et al., 2007) used: ● all or notAll (primary item – hypercube), ● hypercube-dimension, ● dimension-domain, ● domain-member. Fig. 2: Dependencies in XDT (PIECHOCKI et al., 2007) Figure above represents the usage of the taxonomies, elements and connections described by (PIECHOCKI et al., 2007). In case of explicit dimensions all domain members are known and grouped into exactly one dimension. Typed dimensions are used, if the amount of members is too large so that it cannot be expressed with a finished number of members. The domain members within explicit dimensions are connected using the arcrole domain-member creating dimensional hierarchies. Further the explicit dimensions are connected to the domain members via dimension-domain arcrole. Both explicit and typed dimensions are gathered in hypercubes using the hypercube-dimension arcroles (HOFFMAN, 2006). Finally the arcroles all dimensions of a hypercube are excluded from the primary item (HERNANDEZ-ROS et al. 2006). Due to the reason that not each primary item has to be linked to the hypercube, the arcrole domain-member can be used not only within domain member’s taxonomies, but also within the primary taxonomies. This offers the possibility to link a full tree hierarchy of primary items to the respective hypercube. The evaluation of the XDT Modeling Technique divides (PIECHOCKI et al., 2007) to three stages. The highest stage (+) informs about the complete fulfillment of the analyzed criteria. In case the criteria fulfillment level is insufficient the middle stage (o) is assigned. The third stage (-) represents the situation when the criteria is not fulfilled or not concerned. The following figure 6 illustrates the evaluation of the used modeling techniques. Fig. 3: Evaluation result (PIECHOCKI et al., 2007) To summarize the results we can follow the research by (PIECHOCKI et al., 2007) so the multidimensional data can be modeled by using XDT. This is shown by the fulfilled evaluation criteria. Due to the graphically representation of the model elements, data warehouse engineers have an improved understanding of the multidimensional data of this approach because the model elements have more comprehensive semantics. Thus the main advantage is the possibility of mapping between the modeled XBRL taxonomies and the data warehouse schemas. However, XBRL technology should be seen as a complement rather than a replacement for traditional data warehouse/mart-driven BI reporting. Vendors support for XBRL are also growing; many vendors have committed themselves to supporting XBRL as an interchange format for importing and exporting data from their systems. Microsoft's FRX financial software now supports XBRL as a widely accepted format for reporting on and publishing financial information. SAP was one of the early joiners of an international project committee set up to launch XBRL back in 2000. Oracle expanded XBRL support (via a XBRL Manager component) in Enterprise Performance Management System. Enterprise Engineering has released a suite of XBRL-based financial management, analysis and reporting products built on its EnterpriseFTX platform (SHEINA, 2008). 3 Results Different path take us down Chamoni (2007). He analyses the interrelationship between XBRL and business intelligence concepts. Both concepts have common the support and automation of the management process of reporting and analysing business information. Difference is that XBRL tries to describe the meaning of business data and to standardize data exchange; BI seeks to analyse and report these decision-relevant data. Both concepts come from different perspectives, XBRL from semantic description of data within an XML environment and BI from search of knowledge in data. Fig. 4: Proposed processing architecture (HODINKA, 2013) So it is not surprising that an integration of both architectures and specially the convergence of taxonomies will bring benefits to business applications. An architectural design presented on figure 4 in this chapter is more deeply described in dissertation thesis of the first author (HODINKA, 2013) of this article. The three components are the XBRL Processor, XBRL Gate and the Business Intelligence ETL/Replication component. To completely cover any amount of data (including the Big Data), each of the component must by multi-tenant and highly scalable. This design supports the Cloud environment as well as the Private Cloud primarily. Fig. 5: Example of an XBRL report transformation (DEBRECENY, 2013) Even these concepts are founded on similar basis one of the main differences is that XBRL instances are normally snapshots of single data points whereas fact tables in BI systems represent time series. At Chamoni (2007) time, there was only evidence for integration between BI and XBRL. Taxonomies were built in the data layer and used in reporting systems in the presentation layer, more over; our concept includes a separate ETL or replication resource for full adoption on Big Data domain. The connection to the domain of Big Data and XBRL we see through the global reporting services. Now you can see companies in SMB that already acts on global markets and are data driven. Then, the complete automation of the extraction process with minimizing the effort on report data transformation is more then important. And here we see the big additional value from XBRL. The language concept and its adoption from big technological companies show possibility to be widely adopted what confirms the evolution process of the XBRL. Its Formulas and easy translation on multi-dimension models (see figures 5 and 6) have solid potential to be early adopted in the ETL, or ELT, process without any additional investment. Together with the business rules that can easily describe the business logic, which is now generally covered on the application level, makes the whole thing even more powerful not only in the business world, but also in government area like in environmental reporting. Fig. 6: Report-Dimensions mapping (DEBRECENY, 2009) Our aim is to explain how to adopt environmental reporting in Community Environmental Management and Audit Scheme (EMAS) and XBRL-tagged reporting formats. We used Global Reporting Initiative (GRI) industry-specific categorization scheme that defines and “tags” data in relation to its purpose, framework or outline. The key is to identify individual detailed reporting elements, which can be easily shared. Of course, XBRL was not designed explicitly as a BI technology. It was designed as a metadata representation language, but organizations can benefit from well-defined structured format collecting and disseminating sustainability information. With Chamoni approach we can see interesting maturity model for BI where he portrays XBRL playing a native role in areas such as text mining and web reporting. We agree with Debreceny (2007) that this was only important first step, the study by Chamoni provides only a tantalizing preview of future XBRL-based BI implementation and there a clear need for case studies and research pilots that would test the propositions made by Chamoni. 4 Discussion Even XBRL might seem like a finance-only play, but the data exchange standard is flexible enough to support the reporting requirements outside the office of finance. We discussed the relationship between environmental and sustainability indicators and corporate sustainability reporting in chapter “Sustainability indicators: development and application for the agriculture sector” printed by Springer (HŘEBÍČEK et al., 2013). Our research contains the possibility of the utilization of information and communication technology and XBRL taxonomy. We suggest the formalization of the reporting systems of agriculture organizations on the basis of the universal markup language XML by means of the use of the XBRL to minimize main barriers why agriculture organizations do not support sustainability reporting Hřebíček (2009) for example: ● Collecting and managing data is expensive, technical issues with data collection are also a problem. ● Determining the set of appropriate sustainability indicators to monitor and measure is difficult. ● There are difficulties in capturing reliable data information (some aspects of the agrosystem are very difficult to collect meaningful and repeatable data). ● Information disclosure can create business risks, which competitors and regulators may seize upon. ● Difficulty to determine the sphere of influence of an organization. ● Many organizations have good intentions, but simply have not allocated enough resources due to the current economic situation in the Czech Republic. ● Reporting is seen as a superfluous and burdening activity. The core of these barriers is the certain time-demanding nature of the agriculture farm data processing, and the absence of positive feedback (HODINKA et al. 2012). By implementing the XML scheme, the agriculture organization gains a whole set of advantages. The administration and editing of information is much easier and much more effective. Employing the above-mentioned framework enables an improved communication and collaboration with target groups and concerned parties. By implementing the scheme the company acquires the possibility of creating and publishing compact, focused messages that are generated automatically on the basis of the template rules of one single scheme. (HŘEBÍČEK at al., 2013) 5 Summary As we provided in the text of this paper he critical issue in data processing and data analysis tasks is to get the right information quickly, near to real time, targeted, and effectively. If XBRL as data exchange format will be adopted in the whole information supply chain process it will be eliminating most of the costly, often manual, processes of data preparation. This is because XBRL data allows itself to be transformed by software or other mapping tools automatically, which in turn increases consistency, accuracy and trust in data - all key tenets of successful BI reporting. We can see XBRL taxonomies as s start point to build a global data warehouse kernel of qualitative information throughout international corporations. 6 Acknowledgements The Czech Science Foundation supports this paper. Name of the Project: Construction of Methods for Multi Factor Assessment of Company Complex Performance in Selected Sectors. Registration No. P403/11/2085. 7 References [1] DEBRECENY, R., et all. XBRL for Interactive Data. Introduction to XBRL. [http://dx.doi.org/10.1007/978-3-642-01437-6_3]. Springer Berlin Heidelberg 2009, pp. 35-50. ISBN 978-3642014369. [2] DEBRECENY, R., FELDEN, C., PIECHOCKI, M. New Dimensions of Business Reporting and XBRL. Springer, 2007, 271 p., ISBN 978-3835008359. [3] DEBRECENY, R. New Dimensions of Business Reporting and XBRL: Research into XBRL – Old and New Challanges. Springer, 2007, pp. 5-13, ISBN 978-3835008359. [4] GLUCHOWSKI, P., KEMPER, H.-G. Quo Vadis Business Intelligence, In: BI- Spektrum, TDWI, 2006, pp. 12-19. [5] HERNANDEZ-ROS, I., WALLIS, H. (2006). XBRL Dimensions 1.0 Candidate Recommendation,. XBRL International (2006) [online] Filetype RTF. [cit. 2013-01-10]. http://www.xbrl.org/Specification/XDT-CR3-2006-04-26.rtf. [6] HŘEBÍČEK, J. et all. Sustainability Indicators: Development and Application for the Agriculture Sector. Springer Verlag: Sustainability Appraisal: Quantitative Methods and Mathematical Techniques for Environmental Performance Evaluation, 2013, pp. 63-102, ISBN 978-3642320804. [7] HŘEBÍČEK, J., HÁJEK, M., CHVÁTALOVÁ, Z., RITSCHELOVÁ, I. Current Trends in Sustainability Reporting in the Czech Republic. In EnviroInfo 2009. Environmental Informatics and Industrial Environmental Protection: Concepts, Methods and Tools. 23. International Conference on Informatics for Environmental Protection. Aachen: Shaker Verlag, 2009. pp. 233-240, ISBN 978-3832283971. [8] HODINKA, M. Formalization of Business Rules and Methodology of their Implementation to the Information System. Ph.D. thesis. Brno, 2013. [9] HODINKA, M., ŠTENCL, M., HŘEBÍČEK, J., TRENZ, O. Current trends of corporate performance reporting tools and methodology design of multifactor measurement of company overall performance. Acta univ.agric. et silvic. Mendel. Brun., Brno: MZLU, 2012, LX, č. 2, pp. 85–90. ISSN 1211-8516. [10] HOFFMANN, C. Financial Reporting Using XBRL: IFRS and US GAAP Edition. Lulu.com, 2006, 515 p., ISBN 978-1411679795. [11] CHAMONI, P. New Dimensions of Business Reporting and XBRL: XBRL and Business Intelligence – From business Reporting to Advanced Analysis. Springer, 2007, pp. 179188, ISBN 978-3835008359. [12] KLEMENT, T. New Dimensions of Business Reporting and XBRL: Standardized Company Reporting with XBRL. Springer, 2007, pp. 249-271, ISBN 978-3835008359. [13] PIECHOCKI, M., FELDEN, C., GRANING, A. Multidimensional XBRL reporting. In: Proceedings of the Fifteenth European Conference on Information Systems, ECIS, Switzerland, 2007. [14] SHEINA, M. MarketWatch: XBRL bubbling up in BI. (2008). [online] Filetype HTML. [cit. 2013-01-10]. http://netdotwork.co.za/article.aspx?pklarticleid=4965. Sustainability indicators for measuring SME performance and ICT tools Edward Kasem, Jiří Hřebíček Mendel University, Department of informatics, Faculty of Business and Economics Zemědělská 1, 613 00 Brno, Czech Republic xkasem@node.mendelu.cz hrebicek@iba.muni.cz Abstract Small and medium enterprises (SMEs) have significant effects on Europe economy, especially on Czech Republic economy. In this paper, we concentrate on the agriculture SME’s more than others. At the beginning, we describe SMEs impacts on the economy. Then two types of frameworks for measuring and reporting on four pillars of sustainability (economic, environmental, social and governance) are introduced. These frameworks include Global Reporting Initiative GRI and Sustainable assessment of food and agriculture systems (SAFA). Finally, the current state of sustainability indicators and information and communication technologies for measuring complex enterprise performance are presented. Abstract Malé a střední podniky (MSP) mají významný vliv na evropskou ekonomiku, zejména na ekonomiku České republiky. V tomto dokumentu se zaměřujeme na zemědělství SME víc než ostatní. Na začátku, popisujeme MSP dopady na ekonomiku. Poté jsou představeny dva typy rámců pro měření a hlášení o čtyřech pilířích udržitelnosti (ekonomické, ekologické, sociální a vládnutí). Tyto rámce patří Global Reporting Initiative GRI a udržitelnému hodnocení potravin a systémů zemědělství (SAFA). Nakonec jsou uvedeny aktuální stavy ukazatele udržitelnosti a informačních a komunikačních technologií pro měření výkonu v areálu podniku. Keywords GRI, SAFA, SME, sustainability Klíčová slova GRI, SAFA, SME, udržitelnost 1 Introduction Sustainability reporting is a global trend that engages companies in disclosure of their overall economic, environmental, social, and governance impacts and efforts [1]. At the current stage, primarily the large companies only who got involved in sustainability reporting practices. Small and medium sized enterprises (SMEs), despite of their significant impacts in the European economy, employment and environment, show low level of engagement. Because of the important effects of SMEs that impact on the Europe economy, small and medium enterprises need a method or procedure to measure, control and improve their performance. Such a procedure for effective and efficient management should be simple, efficient and integrate these various viewpoints of the performance in environmental, economic and social terms [2]. The main goal of this paper can be abbreviated in assessment the current state (state-of-art) of ICTs for measuring enterprise performance and sustainability of the selected sectors of the national economy. In this paper, we summarized the results that are achieved by the project No.P403/11/2085 [5] in chosen organizations of the agriculture and food processing sector [3], [6]. These results related to ESG aspects of corporate performance evaluation, indicators and reporting. The most of these researches depend on Global Reporting Initiative (GRI G3.1) and Sustainability Assessment of Food and Agriculture (SAFA) systems. Our goal conclude in making the assessment of the current state (state-of-art) of ICTs for measuring enterprise performance and sustainability, then develop it using more advanced version of Global Reporting Initiative like GRI G4. 2 Small and Medium enterprise’s Sustainability The small and medium enterprise (SME) sector is a driver of the business domain, growth, innovation as well as competitiveness. SMEs play a major role in the economic and social dimensions generating the most significant economic value and employment rate in Europe. 25 50 75 100 CZ PT IT ES SE HU NO LU SI EU-27 FI DK FR LT PL AT EE BG NL LV UK DE RO SK Figure 1: Describes the density of SMEs (number of SMEs per 1 000 inhabitants). At the same time, small and medium enterprises are highly sensitive to the quality of the business environment. They have a weaker financial background, because the main goal of their existence is increasing their profit. The importance of small and medium enterprises in the economy is not challenged by anyone. SMEs are estimated to generate 58.6% of the value added within the non-financial business economy. Furthermore, they have a significant contribution to the employment in the EU hiring 66.7% of the non-financial business economy workforce [7]. Figure 1 depicts the density of SMEs in different Europe countries [7]. In Czech Republic, according to 31st December 2010 statistics, there were a total of 1 029 871 small and medium enterprises, representing 99.83 % of all economically active entities. They had a 62.33 % share in the total employment rate and a 36.22 % share in the GDP. A detailed view of the small and medium enterprise development in the Czech Republic in 2000 - 2010 is provided in Figure 2 [8]. Figure 2: Development of the number of active SME in the CR According to the survey, 57% of companies include a reference to sustainable development, and include it into their strategic goals, and 43% of companies do not include a reference to sustainable development and it is not a part of their strategic goals, either. While, sustainability development is an important issue that should be taken into consideration regardless the field of interest of certain organization. Table 1 describes the most frequently practical and effected fields in sustainability reporting according to the conducted SMEs research [8]. Definitely Not Rather Not Rather Yes Definitely Yes In environme ntal field 5.6% 6.4% 16.9% 13.7% In economic field 3.2% 6.5% 15.3% 21.0% In social field 4.8% 7.3% 16.9% 12.9% Table 1: Link to the different areas of sustainability development. 3 Global Reporting Initiative (GRI) Nowadays, the Global Reporting Initiative (GRI) [9], [10], [11] is the most common non-profit organization for developing comprehensive sustainability reporting framework that is widely used across the world. Its mission is to provide a credible transparent framework for sustainability reporting that can be used for all organizations regardless of their size, sector or location. This Framework enables all organizations to measure and report on four key areas of sustainability (economic, environmental, social and governance impacts). GRI is a multistakeholder, network-based organization. Their vision is a sustainable global economy where organizations manage their economic, environmental, social, and governance performance, impacts responsibly, and report transparently. To achieve these goals, they made sustainability reporting standard practice by providing guidance and support to organizations. G3.1 Guidelines [11] was considered as a starting point of the G4 Guidelines [9], [10]. Some changes were made to develop G3.1 and get more qualified Guidelines. The G4 Guidelines are presented in two parts: Reporting Principles and Standard Disclosures [9], and Implementation Manual [10]. Reporting Principles and Standard Disclosures explains the requirements of reporting against the framework, ‘what’ must be reported. Whereas Implementation Manual provides further guidance on ‘how’ organizations can report against G4 criteria. The objectives of developing the next generation of its reporting Guidelines (G4) are to offer user-friendly guidance helps the reporters understand and use these guidelines, improve the technical quality of the guidelines content to eliminate ambiguities and differing interpretations, harmonize with other international accepted standards, and offer guidance related to link the sustainability reporting process to the preparation of an Integrated Report [2]. These changes can be summarized in five most significant areas: Materiality, which takes central stage. Reporting boundaries. Application levels replaced by the concept “In Accordance”. Governance disclosure. Supply chain. 3.1 Materiality Materiality is one of the most important concepts in G4 guidelines. G4 concentrate on the materiality of each aspect. Materiality is not a new concept; it exists in G3 guidelines with the same definition and principle. G4 contains new requirements for organizations to explain the process, which is used to identify their Material Aspects. G4 encourages reporters to focus on content which effects on their business rather than reporting on everything. To begin the process of defining the content of a report, the organization is required to select material Aspects. That means that the materiality determines what sustainability indicators will be reported. This is the major difference between G4 and G3.1, where in G3.1 the reports should be made using certain number of indicator which is called “core indicator”, regardless of whether they were impacted your organization or not. According to all above, GRI G4 describes the steps that the organization may go through in order to define the materiality and the specific content of the report. These steps can be described in Figure 3 and summarized in identification, prioritization, validation and review. More details can be found in G4’s Implementation Manual from pages 32-40. Figure 3: Defining material Aspect and Boundaries process overview 3.2 Reporting boundaries After determine the material aspects, the reporter must consider whether the impact of each issue lies inside, outside, or in the both sides of the organization and describe where the impact ends. To determine whether an impact is within, outside or both, we have to consider what makes the Aspect relevant. For example, material aspects could be relevant primarily to entities inside the organization like anti-corruption. It could be relevant primarily to entities outside the organization, such as suppliers, distributors, or consumers. On the other hand emission could be an aspect, which impacts within and outside of the organization. These what have been extended in G4 to encourage companies to consider and report on a broader range of impacts compared with G3.1 which reports only on impacts that the company has control over or significant influence on. 3.3 The application levels replaced by the concept “In Accordance” In G3.1 Guidelines, there is something called “Application level” system. It appears to declare the level of the reports, which is applied according to the GRI Reporting Framework. To meet the needs of beginners, advanced reporters, and those somewhere in between, there are three levels in the system. These levels are titled C, B, and A and reflect a measure of the extent of application or coverage of the GRI Reporting Framework. A “plus” (+) is available at each level (ex., C+, B+, A+) if the reports utilized external assurance [11]. Whereas the Guidelines G4 offer two options for an organization in order to prepare its sustainability report ‘in accordance’ with the Guidelines: the Core option and the Comprehensive option. The Core option contains the essential elements of a sustainability report. It provides the background against which an organization communicates the impacts of its economic, environmental and social and governance performance. The Comprehensive option builds on the Core option by requiring additional Standard Disclosures of the organization’s strategy and analysis, governance, and ethics and integrity. In addition, the organization is required to communicate its performance more extensively by reporting all indicators related to identified material Aspects. 3.4 New governance disclosure requirements In G4 Guidelines, number of changes has been proposed on governance, along with a new category of disclosure on ‘Ethics and integrity’. These changes are related to information on the composition, involvement and authority of the reporting organization’s highest governance body. They were made to strengthen the link between governance and sustainability performance, taking into account the consistency within existing governance frameworks and developments in that field. The G4 Guidelines contain 10 new standard disclosures on governance (G4.35, G4.36, G4.42, G4.43, G4.46, G4.48, G4.50, G4.52, G4.54, G4.55) [10]. These changes are so important for reporters in order to achieve the ‘Comprehensive’ level of reporting. 3.5 Supply chain More prominences are given for supply chain included new and amended disclosures. Supply chain and supplier are redefined in G4 Guidelines. Companies must disclose how they manage environmental and social issues related to the material impacts of their supply chains. More comprehensive reports should be done related to supply chain including details of supply chain assessments, risks identified, the organization’s performance in managing these risks and the management processes put in place. In additional, the reporter should concentrate on suppliers boundaries and reporting by region, total spend, type of supply, etc. 4 SAFA Sustainability Assessment of Food and Agriculture systems SAFA is a holistic global framework for the assessment of sustainability along food and agriculture value chains [12]. It is one of the most important concept that should be taken into consideration when informing the society about the progress done in the economic, social and environmental and governance areas. This framework can be applied on small, medium and large-scale companies, organizations and other stakeholders that participate in crop, livestock, forestry and fishery value chains. SAFA seeks to harmonize long-term objective of sustainable transformation of food systems, providing a transparent and aggregated framework for assessing sustainability. The guiding vision of SAFA is that food and agriculture systems worldwide are characterized by four dimensions of sustainability; good governance, environmental integrity, economic resilience and social well-being. SAFA Guidelines include 21 themes and 56 sub-themes defined through expert consultations [13]. The Guidelines are the result of an iterative process, built on the cross- comparisons of codes of practice, corporate reporting, standards, indicators and other technical protocols currently used by private sector, governments, not-for-profits and multi-stakeholder organizations that reference or implement sustainability tools. 4.1 SAFA System The guiding vision of SAFA is that food and agriculture systems worldwide are characterized by all four dimensions of sustainability; good governance, environmental integrity, economic resilience and social well-being [12], [13]. SAFA system consists of four nested levels to enhance coherence between each other (see Figure 4) [12]. These four levels begins with SAFA Framework, the highest level, which encompass many aspects by aggregating four dimensions of sustainability; good governance, environmental integrity, economic resilience and social well-being. The second level contains a set of 21 core sustainability issues or universal “themes”. At this level, policy-makers and national governments can work towards alignment and harmonization of a holistic scope of sustainability without defining the specific pathways. In the third level, each of the 21 sustainability themes is divided into sub-themes, or individual issues within SAFA themes. This level composed of 56 sub-themes, is relevant for supply chain actors doing a contextualization analysis, which identifies hot spot areas, as well as gaps in existing sustainability efforts. SAFA Framework Themes(21) Universal Sub-Themes(56) Specific to supply chain Core Indicators(113) For crops, livestock, forestry & fisheries Figure 4: depicts the SAFA System. The last level defines Core Indicators within each sub-theme. These indicators identify the measurable criteria for sustainable performance for the sub-theme. Core Indicators provide ratings for the highest performance (green) and unacceptable practices (red) (This is detailed in SAFA Guidelines Section 2) [12]. 4.2 SAFA Principles SAFA divides into two types of principles: Methodological principles compose of many concepts like holistic, relevance, rigor, efficiency, performance orientation, transparency, adaptability, and continuous improvement. Implementation principles, which describe the following facts: build on existing tools, take place in an open and learning system, and accessibility. 4.3 Scope of SAFA assessment SAFA can be adapted to different contexts and scopes. These scopes can be presented as Supply chain scope starts with the production of a primary commodity, ends with the consumption of the final product and it includes all of the economic activities undertaken between these phases, such as: processing, delivery, wholesaling and retailing. In other word supply chain scope specified in setting the boundaries. Temporal scope which is defining the time frame. For establishing the thresholds of the ratings, especially in the environmental dimension, entity’s activities for one year interval is assumed. But there are some indicators, multi- year trends should be assessed or sustainability impacts be allocated to a longer period – usually in these cases a period of five years is suggested. Thematic scope which is defining the sustainability context. SAFA outlines essential elements of sustainability through 21 high level themes. These are applicable at any level of development, for instance national level or commodity specific. The themes are further divided into 56 sub-themes. SAFA sub-themes are tailored to food and agriculture supply chains and thus, are not well suitable for policy development. Then each sub-theme consists of core indicators proposed in order to facilitate measuring progress towards sustainability. 5 Current state of sustainability reporting according to GRI G3.1 and SAFA guidelines The performance indicators have been developed last few years [1], [2], [14]. In this section, we will depict the latest version of key generic performance indicators achieved by GACR 403 project [5]. This set of indicators is listed in the Table 2 [13] and can be used for wide range of SMEs on the food and agriculture sector [3], [6]. Measurable Performance Indicator Scale and (Unit) Environmental indicators EN01 – Environmental protection investments Investment EN02 – Environmental protection expenditures Emission Resource Consumption Total investments to environmental protection / value added [%] Cost of Environmental protection expenditures /value added [%] EN03 – Total air emissions Total direct and indirect emissions to air / value added [t / CZK] EN04 – Total greenhouse gas emissions Direct and indirect emissions / value EN05 – Total annual energy consumption Direct and indirect energy consumed / value added [MWh /CZK] EN06 – Total renewable energy consumption Energy from renewable resources * 100 / total energy consumption [%] EN07 – Consumed materials added [t / CZK] Consumption of materials by weight / value added [t / CZK] EN08 – Recycled input materials The share of raw materials from recycled materials, percentage of total input materials [%] EN09 – Total annual water consumption The total volume of water removed /value added. [m³ / CZK] EN10 – Total annual waste production The total amount of each type of waste generated by operation of the company / value added [t / CZK] Waste EN11 – Total annual production of hazardous waste The total amount of hazardous waste generated by operation of the company / value added [t / CZK] Social indicators Company SO01 – Community Social investment to local communities * 100 / value added [%] SO02 – Contributions to municipalities Monetary value of projects of municipalities * 100 / value added [%] HR01 – Discrimination Number of discriminatory cases * 100 / number of employees [%] Human rights HR02 – Equal opportunities Labor relations LA01 – The rate of staff turnover Terminated employment relationships * 100 / average number of employees [%] LA02 – Expenditure on education and training Total expenditure on education * 100 /value added [%] LA03 – Occupational Diseases Occupational illness * 100 / average number of employees [%] LA04 – Deaths Number of fatal accidents * 100 / average number of employees [%] PR01 – Customer Loyalty Number of customers at the end of the year – the newcomers during the year * 100 number of customers at the beginning of the year [%] PR02 – Marketing Communications Communications – website, etc. [y/n] PR03 – Health and safety of customers The total monetary value for noncompliance * 100 / value [%] Responsibility for products Number of women * 100 / average number of employees [%] Corporate Governance (CG) indicators Monitoring CG01 – Company and reporting Information Effectiveness of CG Information about the objectives and strategy of the company [y/n] CG02 – CG Responsibility Collective agreement [y/n] CG03 – CG Standardization Publishing of standardized (GRI,vCSR) reports [y/n] CG04 – Ethical behavior Code of ethics [y/n] CG05 – CG Codex Code of corporate governance [y/n] CG Structure CG06 – CG Remuneration Amount of remuneration of board of directors and the supervisory board * 100 / Annual labor costs [%] CG07 – CG Membership Number of independent members of the CG * 100 / number of top management members [%] CG08 – Equal opportunities Number of women in the total number of members of CG [%] CG09 – Compliance with regulatory standards Monetary value of fines for noncompliance * 100 / value added [%] Economic indicators Value added Market position Efficiency EE01 – Value added per employee value added * 100 / average number of employees in the year [%] EE02 – Value added to personnel costs value added * 100 / payroll costs [%] EE03 – Market Share turnover size in manufacturing * 100 / turnover [%] EE04 – Profit EAT Earnings after taxes – EAT [CZK] EE05 – Profit EBT Earnings before taxes – EBT [CZK] EE06 – Profit EBIT Earnings before interest and taxes – EBIT [CZK] EE07 – ROE Performance ROE = EAT / Equity [CZK] EE08 – ROA Performance Return on assets ROA = EBIT / Total assets [CZK] EE09 – ROS Performance Return on sales ROS = EAT / Total sales [CZK] EE10 – ROI Performance Return on invested capital ROI = EBIT / Total equity [CZK] EE11 – ROCE Performance ROCE = EBIT / Equity + Long-term liabilities [CZK] EE12 – Turnover size Sales of own products and services * 100 / value added [%] Cash Flow EE13 – Operating Cash Flow Total cash resources of the company * 100 / Total operating costs [%] EE14 – Expenses on R &D Total costs * 100 / value added [%] EE15 – Number of employees Average number of employees in the year (persons) [number] Table 2 ESG indicators proposed for manufacturing enterprises Also a prototype of reporting information system was implemented. This prototype was made to insure that the indicators can be implemented successfully and verify usability of information system for end user. It concentrates on assessing the sustainability of crop production systems for the conditions of the Czech Republic. Figure 5 depicts the core data model of the reporting system. According to the figure 5 the basic indicator abstraction entities are: indicator group, indicator, indicator item, indicator item aspect. This prototype describes a developed infrastructure for ICT tools [15] to support rapid knowledge-based decision making for sustainable development and integrated information from various environmental social and corporate governance impacts related with economy [1][2]. More information about this prototype can be found in [13] [16]. 6 XBRL XBRL, eXtensible Business Reporting Language [17], is an XML-based markup language used to communicate financial and business data electronically. It is generally user friendly because it isn’t requiring any previous XML or IT background. XBRL is used to encode financial information such as financial statements, earnings releases, and so forth. Using XBRL, the financial information can be automatically and more easily sorted and compared. Because computers don’t have any knowledge about financial reporting and do not understand information that is not fully defined, XBRL making the information machine-readable. companyid Associated Indicator companyid IndicatorReportType Report Type PK PK,FK1 indicatorid PK,FK2 reportTypeid FK3 IndicatorReportType reportTypeid name version Indicator PK PK,FK1 parentIndicatorid PK,FK2 childIndicatorid indicatorid name formula FK1 indicatorGroupid description Indicator group PK Indicatorgroupid name IndicatorReportGroup Company PK indicatorReportGroupid PK name name FK1 parentid Report Instance companyid UserCompany PK PK userid companyid PK reportInstenceid companyid reportTypeid FK3 period Indicator Value PK indicatorValueid FK3 period FK1 indicatorItemid FK2 indicatorAspectid value FK4 dataTypeid Company Period PK FK1 companyid name userid Data Type Indicator Item PK dataTypeid PK indicatorItemid name formula FK2 dataTypeid FK1 indicatorid description Indicator Item Aspect PK indicatorItemAspectid FK1 indicatorItemid formula FK2 dataTypeid description PK periodid Figure 5: describes the core data model of the reporting system. In other words, XBRL is a tool that benefits all users of financial statements by providing increased functionality over traditional paper, Hyper Text Markup Language (HTML), and other image-based financial reporting formats. The markup codes used in XBRL describe financial data in a format that computers can classify, sort, and analyze. The analysis of these financial data will be done using pre-existing names to standardize business-reporting terminology. For example different companies usually use different terms to describe the same concept (accounts receivable, receivables, and trade [accounts] receivable). XBRL normalizes company-specific terminology to get to the root business reporting concepts, allowing comparison of essentially similar information across multiple financial statements and among multiple companies. XBRL doesn’t use for change how, what, or when to report financial information or change financial reporting disclosure requirements. It simply provides a standard mechanism for computers to understand and compare financial data by transmitting financial information in a way that leverages computer technology. XBRL assigns appearance and organizational characteristics to electronic documents which can be used in Internet other website content. According to the current standard, XBRL 2.1 [18], published by XBRL International to perform tagging, preparers must have a collection of previously defined elements that provide the definitions that computers use to understand financial statement data. These collections of defined elements are called taxonomies. To create XBRL-encoded financial statements, preparers match each piece of information from the financial statements to an element in the taxonomy. During this matching process, known as mapping, preparers note where the taxonomy differs from the financial statements. In such cases, preparers extend the XBRL US GAAP Taxonomy by relabeling elements, rearranging elements, or creating new elements to match the financial statements. Once preparers have a fully mapped financial statement and extend the XBRL US GAAP Taxonomy to match the financial statements, they can then tag each piece of financial statement information to create an XBRL-encoded financial statement known as an instance document. 7 Conclusion In this paper, we make an assessment of the current state (state-of-art) of the sustainability indicators and information and communication technologies (ICTs) which is used for measuring enterprise performance and guarantee its sustainability of the food and agriculture sectors of the UE economy from theoretical and practical point of view. All explained indicators depend on the SAFA and GRI 3.1 guidelines. These days a new GRI G4 Guidelines have been lunched. So we present a brief overview about the new GRI G4 and explain its new requirements. SAFA system with its levels, principles and scope are explained. Extensible Business Reporting Language (XBRL), which is used to compare financial data and make it understandable for computer, is mentioned. Future work will concentrate on designing the economic indicators of performance in relation to environmental, social and corporate governance indicators according to new guidelines GRI G4. Develop advanced methods, which identify key performance indicators for ESG performance of chosen company with the possibility of the utilization of information and communication technology and XBRL taxonomy for corporate sustainability reporting. Develop an ICT infrastructure powerful ICT tools to support rapid knowledge-based decision making for sustainable development and integrated information from various environmental social and corporate governance impacts related with economy. 8 References [1] Chvátalová, Z., Kocmanová, A., Dočekalová, M., 2011: Corporate Sustainability Reporting and Measuring Corporate Performance, In Environmental Software Systems. Frameworks of eEnvironment. 9th IFIP WG 5.11 International Symposium. ISESS 2011. Heidelberg: Springer, 398–406, ISBN 978-3-642-22284-9. [2] Hřebíček, Soukopova J., Štencl, M., Trenz, O. 2011: Integration of Economic, Environmental, Social and Corporate Governance Performance and Reporting in Enterprises. Acta Universitatis Agriculturae et Silviculturae Mendelianae Brunensis 59, 7: 157–177, ISSN 1211-8516. [3] Sustainability Reporting Guidelines: Construction and Real Estate Sector Supplement based on GRI G3.1, [online] Available: https://www.globalreporting.org/reporting/sector-guidance/sector-guidance. [4] ISO 9000 ISO 14000 ISO 18000 documentation. [online] http://www.isohelpline .com/ims_iso_9000_iso_14000_ohsas_18000_document_manual.htm Available: [5] Project GACR P403/11/2085, Construction of Methods for Multifactor Assessment of Company complex performance in selected sector. [online] Available: http://www.gacr403 .cz/en/ [6] J. Hřebíček, O. Popelka, M. Štencl, O. Trenz (2012) Corporate performance indicators for agriculture and food processing sector. Acta universitatis agriculturae ET silviculturae mendelianae brunensis. [7] European Business: figure and facts: Eurostat statistical books (2009). [online] Available: http://epp.eurostat.ec.europa.eu/portal/page/portal/european_business/publications/fac ts_figures [8] Kocmanova, M. Docekalova, P. Nemecek, I. Simberova. Sustainability: Environmental, Social and Corporate Governance Performance in Czech SMEs. In: The 15th World Multi-Conference on Systemics, Cybernetics and Informatics. 2011. IFSR, Orlando, USA: WMSCI 2011, 94–99, ISBN 978-1-936338-42-9. [9] G4 Guidelines, 2013: G4 Guidelines: Reporting principles and standard disclosures, [online] Available: https://www.globalreporting.org/reporting/g4/. [10] G4 Guidelines, 2013: G4 sustainable reporting Guidelines: Implementation manual, [online] Available: https://www.globalreporting.org/reporting/g4/. [11] G3.1, 2011: G3.1 Guidelines, http://www.globalreporting.org/ReportingFrame work/G31Guidelines/ Available: [12] SAFA Sustainability Assessment of Food and Agriculture systems. SAFA Guidelines version 2.0, [online] Available: http://www.fao.org/nr/sustainability/sustainabilityassessments-safa/en/ [13] Popelka, O., Hodinka, M., Hřebíček, J., Trenz, O. Information System for Global Sustainability Reporting. In Environmental Software Systems. Fostering Information Sharing, 10th IFIP WG 5.11 International Symposium, ISESS 2013, Neusiedl am See, Austria, October 9-11, 2013. Proceedings. Berlin Heidelberg : Springer, 2013. ISBN 978-3-642-41151-9, pp. 630-640. 9.10.2013, Neusiedl am See, Austria. [14] Hřebíček, Soukopova J., Štencl, M., Trenz, 2011: Corporate Key Performance Indicators for Environmental Management and Reporting. Acta Universitatis Agriculturae et Silviculturae Mendelianae Brunensis 59, 2: 99–108, ISSN 1211-8516. [15] Hodinka, M., Štencl, M., Hřebíček, J., Trenz, O.: Current trends of corporate performance reporting tools and methodology design of multifactor measurement of company overall performance. Acta univ. agric. et silvic. Mendel. Brun., 2012, LX, No. 2, pp. 85–90. [16] Popelka, O., Hodinka, M., Hřebíček, J., Trenz O. Information system for corporate performance assessment and reporting. Proceedings of the 2013 International Conference on Systems, Control, Signal Processing and Informatics (SCSI 2013). : EUROPMENT, 2013. ISBN 978-1-61804-204-0, pp. 247-253. 16.7.2013, Rhodes Island, Greece [17] XBRL, 2012: An Introduction to XBRL. [online] [cit. 2012-24-3] Available: http://www.xbrl.org/ WhatIsXBRL/. [18] Homer, B., Hamscher, W., Westerlund, L., Illuzzi, K.L., Cohen, L.P., and Cohen, D.: XBRL US GAAP Taxonomy Preparers Guide. 2008 XBRL US. Curriculum planning and construction: How to combine outcome-based approach with processoriented paradigm Martin Komenda1,2, Lucie Pekárková2 1 Institute of Biostatistics and Analyses, Kamenice 126/3, 625 00, Brno 2 Faculty of Informatics, Botanická 68a, 602 00, Brno komenda@iba.muni.cz, xpekark@fi.muni.cz Abstract The paper discusses a proposal of innovative combination of two various approaches used during a curriculum planning lifecycle. Outcome-based and process-oriented paradigms are introduced regarding the need for replacing current models in tertiary education. Outcomebased education helps teachers to tell students more precisely what is expected of them and is being implemented in many institutions. On the other side, process-oriented paradigm can strongly enhance the design of curriculum composition. The basic goal of this contribution is to highlight a challenge how to find general concept for planning and construction of the curriculum and course structure, followed by theoretical background of learning. Abstract Příspěvek pojednává o kombinaci dvou různých přístupů používaných během plánování životního cyklu výuky. Představuje procesně orientovaný přístup a přístup založený na výstupech z učení s ohledem na opakovaně akcentovanou potřebu změny současně používaných modelů v terciálním vzdělávání. Přístup založených na výstupech z učení pomáhá pedagogům jasně formulovat, co od studentů vlastně očekávají, a i proto je využíván na mnoha univerzitách. Na druhé straně stojí procesní přístup, který posiluje návrh skladby celého kurikula. Základním cílem příspěvku je v souladu s teoretickými principu procesu učení se zdůraznit nové výzvy v podobě hledání obecného konceptu pro plánování a konstrukci kurikula, a to včetně struktury jednotlivých kurzů, Key words Outcome-based approach, process-based approach, curriculum planning, patterns, process modeling Klíčová slova Přístup založený na výstupech z učení, procesně orientovaný přístup, plánování učiva, vzory, procesní modelování 1 Introduction In recent years, innovation of the curriculum has gained increasing attention especially in the tertiary sphere. The need for a more fundamental paradigm shift in education has been accented. There appears to be some space for replacing the current model of the educational process [1]. An appropriate curriculum structure is required, so universities and other stakeholders have invested a great deal of time and funding into a detailed curriculum review process and subsequent optimization. Curriculum planning and development is very much on today’s agenda for undergraduate, postgraduate and continuing education. The whole curriculum is more than just a syllabus or a statement of content. It is about what should happen in a teaching program – about the intentions of the teachers and about the way they make these happen. This extended vision of a curriculum is illustrated in Fig. 1 [2]. There have appeared a lot of innovative ways how to improve curriculum planning, the following chapter describes in detail method based on learning outcomes. Fig. 1: A wider view of a curriculum [2]. 1.1 Outcome-based approach The new paradigm has introduced outcome-based education, a performance-based approach at the cutting edge of curriculum development, which offers a powerful and appealing way of reforming and managing education. The emphasis is on the product – what sort of university graduates will be produced – rather than on the educational process. In outcome-based education the learning outcomes are clearly and unambiguously specified. These determine the curriculum content and its organization, the teaching methods and strategies, the courses offered, the assessment process, the educational environment and the curriculum timetable. Among the major advantages in adopting an outcome-based model for tertiary education are relevance, controversy, acceptability, provision of a framework, clarity, accountability, selfdirected learning, guidance in assessment, participation in curriculum planning and also a tool for curriculum evaluation [3]. The concept of learning outcomes has been widely used in education research [4]. The paper focuses on an educational area covering tertiary fields of study, so close attention has been paid to experiences in university programs, where there are also demands for defining new aims of innovation, such as: (i) the application of innovative thinking and approaches to the new learning technologies including e-learning and virtual reality; (ii) the application of new approaches to curriculum planning and advanced instructional design incorporating approaches such as curriculum mapping, outcome-based education, the use of electronic study guides, peerto-peer learning, dynamic learning and a bank of “reusable learning objects”; (iii) the provision of a unique international perspective which takes into account the trend towards globalization and offers benefits to society, to universities and to students; (iv) the development of a flexible curriculum which meets the needs of different students [5]. Jenkins and Unwin [6] suggest that learning outcomes help teachers to tell students more precisely what is expected of them and also provide benefit to teachers themselves [7]. Fig. 2: A model for the curriculum emphasizing the importance of educational outcomes in curriculum planning [7]. 1.2 From theory to practice This approach promotes user-friendly methods for planning, implementation and evaluation, especially applied in medical and health care oriented fields of study but not only for that. The integration of learning outcomes into the curriculum can be presented in a number of ways depending on the objectives and goals of individual study programs. For example, Brown University described their learning outcomes as a list of nine abilities, such as effective communication, basic clinical sciences, diagnosis, management, prevention, etc. [8]. The English National Board of Nursing, Midwifery and Health [9] has identified ten key characteristics as the basis for the learning outcomes required for the Higher Award. The Association of American Medical Colleges in the USA has developed a set of goals for medical education. These attributes were memorably stated in four groups: the physician must be altruistic, knowledgeable, skillful, and dutiful [10]. The University of New South Wales has applied the outcome-based approach and has introduced a new curriculum in which the capabilities of graduates are central to its design. In developing its curriculum, the idea of “capabilities” was adopted from the Capability in Higher Education movement in the United Kingdom. Twelve capabilities include not only knowledge and skills but also the capacity to continue to learn from experience, to act in unfamiliar and changing contexts, to be clear on professional purpose and to successfully function with others in the workforce [11]. Schwarz [12] introduced the concept of global minimum essential requirements (GMER), which present a set of global minimum learning outcomes for graduates of medical schools. This approach is produced by the Institute for International Medical Education (IIME) and does not imply a global uniformity of medical curricula and educational processes. Medical schools should adopt their own particular curriculum design, but in so doing they should first assure that their graduates will possess the core competences stated in the GMER document and, second, the competences necessary to meet the unique healthcare needs of the area they serve. Simpson et al [13] described set of learning outcomes that clearly define the abilities of medical graduates from any of the five Scottish medical schools. The outcomes are divided into twelve domains that fit into one of three essential elements for the competent and reflective medical practitioner. They are similar to the seven domains proposed by the IIME. In the Scottish approach more emphasis is given to the relationship between the outcomes and to the integration of the knowledge, skills and attitudes in the practice of medicine. Fig. 3: Desirable capabilities of graduating medical students at the University of New South Wales [11]. Bloch and Bürgi [14] introduced the implementation of the Swiss Catalogue of Learning Objectives for Undergraduate Medical Training, which was inspired by the Dutch model [15] for training doctors based largely on the traditional disciplines. The Dutch blueprint more clearly reflected the discipline-based nature of medical academe in Switzerland. The Swiss learning objectives represent a more discipline-based view of medical education than that found in the IIME or the Scottish outcomes. The first section of the catalogue sets out the profile of the doctor by the end of undergraduate education under four headings. The second section is subdivided into medical aspects, scientific aspects, personal aspects and aspects related to society and the healthcare system, each with several further subheadings. A third section of the report lists 283 problems in medicine. These are defined as a complex of complaints, signs and symptoms, which may lead a patient to seek medical counsel [16]. 2 Challenges and objectives Outcome-based education is being implemented in many institutions despite controversy and debate about whether learning outcomes can be clearly defined. The curriculum represents a multifaceted concept at a variety of levels and the chosen form always depends on individual institution requirements. Obviously a lot of academic institutions across the world have been using the concept of outcome-based curricula in different ways and in various forms. The goal is still the same, to improve the current educational process. The precise reforming of curriculum covers not only sophisticated planning and using of appropriate theoretical approach. We have to take into the consideration the physical format and structure of every individual course, which represents the basic building unit of the common curriculum. There appears a challenge how to find general paradigm for design the curriculum structure, followed by theoretical background of learning and teaching. The second need has been already fulfilled by the use of outcome-based approach. The paper introduces the way of creating a new curriculum arrangement with the use of process-oriented paradigm. 2.1 Process-based approach Curriculum structure is like the skeleton of a body: strong and in balance, giving direction and support to activities and determining the outline of what it represents. It needs careful consideration and planning. The characteristics of the ‘skeleton’ may be determined by the method of teaching, e.g. outcome-based learning [17]. We describe the structured methodology named process-based approach, which can strongly enhance the process of curriculum composition design. This approach provides guidelines which were subsequently used to inform the development of a process-based approach to performance measurement system design [18]. From the tertiary curriculum perspective, it considers every instance of a course as a process. A process is a repetitive sequence of interdependent and linked procedures with the beginning and the end, which, based on one or more inputs, produces demanded outputs. So, the process is a sequence of tasks, which relate to each other and transform inputs (students' and teachers' effort) into desired outputs (students with gained knowledge, their grades or certificates). Of course, there must be a reason to compare courses to processes. When the behavior is defined, e.g. basic checkpoints such as beginning, end, sequence of tasks leading from beginning to end, it can get very predictable and therefore it is simpler to monitor and evaluate processes in the course. That also means that the whole learning process/course itself can be automated and optimized. The introduction of roles, definition of rules, procedures, repeatable processes and some principles of process management can contribute to a comprehensive learning environment. In the business world, the business process management (BPM) implementation leads to one major goal, which is cost reduction and profit improvement. In the e-learning domain the goal could be higher efficiency and quality of education. Implementation and automation of processes in e-learning should project as a combination of direct and non-direct effects [19] that lead the further learning process improvement. Direct effects are: students have clearly defined tasks and duties including deadlines which supports continuous work; students can share information easily due to a knowledge base, which is a space for sharing and storing educational materials; students have instant access to comparison with other students in desired situations for positive motivation. There are non-direct effects: number of administrative tasks are reduced and automated thus this supports team work in the tutor team, which leads to even distribution of duties and competences. The tutor then doesn't need to worry about administration and has more time to prepare the content; the tutor has an instant overview of the situation status, completed and uncompleted tasks thus the tutor team can flexibly react on actual situation; possibility to continuously evaluate efficiency and to optimize the learning process. To find learning processes is a pedagogical problem - every course or lecture needs different approach. A good learning process often involves a lot of activities that need to be done on both tutor’s and student’s side and the complexity of learning process organization is growing and needs to be managed. The most obvious case of a process from the university environment is a course. The course is taught repeatedly and the structure remains mainly the same. Modeling of courses with a lot of mixed e-learning class work, teamwork and other activities is very useful. In case of really rich and interactive course the organization is too complex and starts to take too much of time to prepare it. At this moment the process-oriented approach can help to automate certain tasks and at the same time keep the organization complexity under control. On the other hand, modeling of each process from scratch is time consuming, but in general courses consist of a lot of same sequences of activities (such as teamwork, assignment evaluation etc.). As process approach gives emphasis on re-usability with help of global sub- processes, it is possible to identify certain sub-processes in learning process that can be reused later – the learning patterns [20], [21]. The patterns are used to capture best practices from the educational area. The implementation of mentioned patterns could ease design and optimize learning processes and also increase the quality of education. On the other hand, formalization of patterns through processes can lead to further discussion of the whole learning process. The tutor chooses suitable patterns and eventually compiling single patterns together creates the whole course. Compiled courses have to be functional and make sense likewise from the technical point of view. The educational pattern is a small reusable sequence of steps that are included in one learning activity. Patterns should be used as building blocks for more complex courses [22]. Each pattern should define general standard structure of the activity that can be extended according to specific needs of other (re-)users, i.e. it should be customizable. One of modeled processes is illustrated on Fig. 4. Colloquium is a method of examination, where small group of student discuss gained knowledge with the teacher. Each student also defends elaborated essay. The process starts when the teacher publishes topics for essays, then each student is supposed to elaborate his/her own final essay, submit it, afterwards read essays of two colleagues and write a report which will be discussed at the colloquium. The process ends when teacher submits evaluation. Fig. 4: Example of modeled "Colloquial exam" process This model was created with the project MEDUSY (Multi-Purpose EDUcation SYstem1) that is focused on development of rich e-learning environment that will enable wide integration of various e-learning tools. Project tries to leverage some of the e-learning best practices used in modern learning management systems (LMSs) and introduce process-oriented approach to course development and education flow in general. One functional prototype of process based e-learning application, tool for learning process modeling and basic stones of pattern repository have been created. Also a prototype of the engine was designed which enabled to analyze possibilities of process management and patterns utilization in e-learning/blended learning. Integrating monitoring and evaluation tools into the project will follow. 1 http://medusy.sf.net/ 3 Conclusions The combination of outcome-based approach along with process-based architecture can significantly help to find a way to plan, respectively restructure, the curriculum. The key fact is that curriculum designers attention is focused not only on formulation of concrete required outputs that make up the minimum level of graduate's knowledge, but also on the design of individual courses. These can be designed in many different ways, but courses always consist of clearly defined building blocks, the learning patterns, that are mutually connected and form a comprehensive learning unit (the course). Due to the wide utility of this methodology, its practical applicability in LMSs is an interesting stimulus for the future. An indisputable advantage of this approach is the possibility of monitoring and subsequent analytical processing of information with a clear purpose - to improve the quality of the tertiary education. Interconnection between two seemingly different approaches provides standardization not only of the learning outcomes, but also of the learning patterns. 4 References [1] R. M. Harden, “Looking back to the future: a message for a new generation of medical educators,” Med. Educ., vol. 45, no. 8, pp. 777–784, 2011. [2] R. M. Harden, “Approaches to curriculum planning,” Med. Educ., vol. 20, no. 5, pp. 458– 466, 1986. [3] R. M. Harden, J. R. Crosby, and M. H. Davis, “AMEE Guide No. 14: Outcome-based education: Part 1ÐAn introduction to outcome-based education,” Med. Teach., vol. 21, no. 1, pp. 7–14, 1999. [4] Y. Mong, M. Chan, and F. K. H. Chan, “Web-based outcome-based teaching and learning - An experience report,” in Advances in Web Based Learning - Icwl 2007, vol. 4823, H. Leung, F. Li, R. Lau, and Q. Li, Eds. Berlin: Springer-Verlag Berlin, 2008, pp. 475–483. [5] R. M. Harden and I. R. Hart, “An international virtual medical school (IVIMEDS): the future for medical education?” Med. Teach., vol. 24, no. 3, pp. 261–267, May 2002. [6] Jenkins and D. Unwin, “Writing learning outcomes for the Core Curriculum,” NCGIA GISCC Learn. Outcomes, 1996. [7] R. M. Harden, “AMEE Guide No. 14: Outcome-based education: Part 1-An introduction to outcome-based education,” Med. Teach., vol. 21, no. 1, pp. 7–14, 1999. [8] Smith S. R. and Dollase R., “AMEE guide No. 14: Outcome-based education: Part 2Planning, implementing and evaluating a competency-based curriculum,” Med. Teach., vol. 21, no. 1, pp. 15–22, 1999. [9] “Ten Key Characteristics of the ENB Higher Award,” ENB Publ., vol. 1991. [10] AAMC, “Guidelines for Medical Schools,” Rep. 1 Learn. Object. ForMedical Stud. Educ., 1998. [11] S. B. Dowton, “Imperatives in medical education and training in response to demands for a sustainable workforce,” Med. J. Aust., vol. 183, no. 11/12, pp. 595–8, Dec. 2005. [12] M. R. Schwarz and A. Wojtczak, “Global minimum essential requirements: a road towards competence-oriented medical education,” Med. Teach., vol. 24, no. 2, pp. 125–129, 2002. [13] “The Scottish doctor—learning outcomes for the medical undergraduate in Scotland” [14] R. Bloch and H. Bürgi, “The Swiss catalogue of learning objectives,” Med. Teach., vol. 24, no. 2, pp. 144–150, 2002. [15] J. C. Metz, A. M. M. Verbeek-Weel, H. J. Huisjes, and A. van de Wiel, “Blueprint 2001: training of doctors in The Netherlands,” Utrecht NFU, 2001. [16] “Developments in outcome-based education.” [17] R. Oliver, H. Kersten, H. Vinkka-Puhakka, G. Alpasan, D. Bearn, I. Cema, E. Delap, P. Dummer, J. P. Goulet, T. Gugushe, E. Jeniati, V. Jerolimov, N. Kotsanos, S. Krifka, G. Levy, M. Neway, T. Ogawa, M. Saag, A. Sidlauskas, U. Skaleric, M. Vervoorn, and D. White, “Curriculum structure: principles and strategy,” Eur. J. Dent. Educ., vol. 12, pp. 74–84, 2008. [18] A. Neely, J. Mills, K. Platts, H. Richards, M. Gregory, M. Bourne, and M. Kennerley, “Performance measurement system design: developing and testing a process-based approach,” Int. J. Oper. Prod. Manag., vol. 20, no. 10, pp. 1119–1145, 2000. [19] D. Tovarňák, “Procesní řízení v e-learningu,” 2011. [20] O. Marjanovic, “Towards A Web-Based Handbook of Generic, Process-Oriented Learning Designs.,” Educ. Technol. Soc., vol. 8, no. 2, pp. 66–82, 2005. [21] T. Kolář and T. Pitner, “Process-based approach to e-learning educational processes and patterns,” in Sborník sedmého ročníku konference o e-learningu – SCO 2011, 2011. [22] P. Avgeriou, A. Papasalouros, S. Retalis, and M. Skordalakis, “Towards a pattern language for learning management systems,” Educ. Technol. Soc., vol. 6, no. 2, pp. 11–24, 2003. Information Technology Tools for Current Corporate Performance Assessment and Reporting Ondřej Popelka, Michal Hodinka, Jiří Hřebíček, Oldřich Trenz Department of Informatics, Faculty of Business and Economics Mendel University Zemědělská 1, 61300 Brno, Czech Republic ondrej.popelka@mendelu.cz, michal.hodinka@mendelu.cz, hrebicek@mendelu.cz, oldrich.trenz@mendelu.cz Abstract Current trends of corporate performance assessment (measurement of economic/financial, environmental, social and governance key performance indicators) and information and communication technologies (ICT) tools for corporate reporting are discussed in the paper. We focus on corporate performance assessment particularly in the agricultural and food processing sector. The aim of the paper is the introduction of an information system prototype proposal for corporate sustainable reporting designed for small and medium-sized enterprises (SME). We describe a relational data model of this information system with several abstract entities to represent miscellaneous differences in various corporate reporting frameworks. Our goal is to design a generic information system of corporate sustainable reporting which may be used by SMEs to start with their corporate performance assessment and reporting. The paper describes the overall structure of the information system, along with the core data model and an example of possible extensions into XBRL-based reporting and business intelligence. Abstrakt Současné trendy hodnocení podnikové výkonnosti (měření ekonomických a finančních, environmentálních, sociálních, správních a řídících klíčových ukazatelů výkonnosti) a nástrojů informačních a komunikačních technologií (ICT) pro podnikový reporting jsou diskutovány v příspěvku. Zaměřujeme se na hodnocení podnikové výkonnosti zejména v odvětví zemědělství a zpracování potravin. Cílem článku je představení návrhu prototypu informačního systému pro podnikový reporting pro malé a střední podniky (MSP). Popisujeme relační datový model tohoto informačního systému s několika abstraktními entitami představují různé rozdíly v různých rámcích podnikových reportů. Naším cílem je navrhnout obecný informační systém pro podnikový reporting, který může být používaný v malých a středních podnicích s jejich hodnocení jejich výkonnosti a vytváření reportů. Clánek popisuje strukturu celého informačního systému, jádro datového modelu a příkladem možného rozšíření na bázi XBRL reportu a využití business intelligence. Keywords Corporate performance, Information system, GRI, XBRL, SAFA, Corporate reporting, Performance assessment, Key performance indicators, Business Inteligence Klíčová slova Podniková výkonnost, informační systém, GRI, XBRL, SAFA, podnikový reporting, hodnocení výkonnosti, klíčové ukazatele výkonnosti, Business Inteligence 1 Introduction The research teams of the Faculty of Business and Management (FBM) of Brno University of Technology (BUT) and the Faculty of Business and Economics (FBE) of Mendel University in Brno (MENDELU) have been taking part in the research project No P403/11/2085 “Construction of Methods for Multi-factorial Assessment of Company Complex Performance in Selected Sectors” since January 2011. The project is being solved through the years 2011-2013 and funded by the Czech Science Foundation. The main research goals of this project are specified by its six partial research targets [6], [10], [11], [15]: The analysis of the state-of-the-art of economic, environmental, social and governance aspects of corporate performance through targeted research of the global information and database sources available at the FBM BUT and the FBE MENDELU using the available Information and Communication Technologies (ICTs) tools. A detailed analysis of the implementation of economic, environmental, social and governance reporting in chosen economic activities and its justification. Assessment, analysis, and the categorization of contemporary characteristics of the individual pillars of corporate performance in relation to the measure of progress or dynamics of the development of overall corporate performance. The identification of the importance and relative roles of Environmental, Social and Governance (ESG) factors using ESG data and Key Performance Indicators (KPIs) in the overall company performance. The construction of quantitative and qualitative methods of the multifactor measurement of corporate performance in the chosen economic activities with the use of ICT tools. The application of the developed methods for multifactor measurement of corporate performance of chosen economic activities in practice, with feedback for possible change correction aimed at further improvement. We will discuss the current state of corporate sustainability reporting following G4 Sustainability Reporting Guidelines [21] and SAFA Guidelines version 2.0 [18] in the paper. Further we introduce a developed prototype of the specialized corporate reporting information system [22] for small and medium-sized enterprises (SME). We will present also two primary goals of the prototype implementation: a verifying that the proposed corporate performance indicators can be implemented successfully and usability of this corporate reporting information system for end-users. Though our main goal is the implementation of the SAFA Guidelines methodology [18], we have currently implemented the methodology SAGROS (Sustainable AGROSystems) [16] for assessing the sustainability of crop production systems for the conditions of the Czech Republic. 2 Corporate Reporting Reporting on sustainability and environmental, social and governance (ESG) performance is a crucial step towards a market that rewards the creation of long-term wealth in a just and sustainable society [30]. Sustainability key performance indicators can play a crucial role in supporting markets that create such long term wealth [9]. They can form the backbone of sustainability disclosure that tracks, and allows for improvement on, those issues most tied to a corporation’s environmental and social impact, and are most material to a company’s financial performance [11], [14], [23]. Nowadays, we can see significant change in how reporting systems are evolving. They are moving from the usage of few analysts to a ubiquitous use across many organizational functions. We focus on the development an information system accessible to small and medium-sized enterprises. Large corporations usually already have some kind of reporting system implemented. SMEs usually perform only the mandatory reporting required by laws, basically in order to fulfill regulatory demands. Our goal is to facilitate performance reporting even for SMEs and make it as easy as possible. This should help enhance the strategic decision making and planning of SMEs, which in turn will help to improve the risk management and sustainability of the business and its environment. The Food and Agriculture Organization of the United Nations (FAO) is developing the Sustainability Assessment of Food and Agriculture systems (SAFA) guidelines [18] to assist the achievement of fair practices in food and agriculture production and trade on a local and regional level. The SAFA framework is the result of an iterative process, built on the cross-comparisons of codes of practice, corporate reporting, various standards, indicators and other technical protocols currently used by food and agricultural enterprises that implement sustainability assessment. The SAFA Guidelines are based on certain core methodological principles including Bellagio Stamp [24]. Additionally, SAFA draws upon the ISO standards for Life Cycle Assessment (ISO 14040:2009), the ISEAL Code of Good Practice [12], [25], the Reference Tools of the Global Social Compliance Program (GSCP) [26] and the GRI Sustainability Reporting Guidelines G3 [4], G3.1 [3] and G4 [21] and its Food Processing Sector Supplement [32]. The guiding vision of SAFA is that food and agriculture systems worldwide be characterized by environmental integrity, economic resilience, social well-being and good governance. In recent years, there has been some progress in the realization and acknowledgement of sustainable development, which is summarized in [5], [7] [8] and [9]. Many stakeholders in the agriculture sector have contributed to this progress by improving agricultural productivity, protecting natural resources and human resources by implementing frameworks and standards for assessing and improving sustainability across the agricultural sector and along the value chain [18]. The SAFA guidelines consist of three core sections – the SAFA framework, the SAFA implementation and the Sustainability dimensions [18]. In the following chapters, we describe a special information system, whose role may be divided into several main steps: collecting data; reporting in standardized format for the purpose of stakeholders and required regulatory demands; reporting in XBRL standardized interchange format; customized reporting; evaluating company performance by computing standardized key performance indicators. These steps correspond to the last two steps of Section 2 of the SAFA framework implementation (indicators, reporting). To cover the first two steps (mapping, contextualization) seems impractical at present due to the highly specific nature of the problem for each reporting SME / corporation. Therefore, these steps are not considered a part of the proposed information system. One of our goals is the development of a special information system or information system module for SME performance reporting. These organizations would be able to use such a system to generate standardized reports and assess their corporate performance. 3 Overview of existing performance Indicators 3.1 SAFA methodology Because we concentrate mainly on agricultural organizations, the SAFA methodology [18] is highly important to us. To allow a greater general scope of the information system for SMEs we have also considered and evaluated other performance methodologies and their indicators. There are several levels of SAFA, which are nested to enhance coherence at all levels (see Figure 1). The SAFA Framework begins with the high level overarching dimensions of sustainability: good governance, environmental integrity, economic resilience and social well-being. It is recognized that these dimensions are broad and encompass many aspects. These are translated into a universally agreed definition of sustainability, through themes and sub-themes for each of the sustainability pillars. Together it is 21 themes and 56 sub-themes. These are measurable and verifiable through a set of 112 core indicators applicable to food and agriculture supply chains. SAFA Guidelines [18] provide the guidance for the application (calculation) of these indicators. Fig. 1. SAFA Guidelines structure [18] SAFA has defined 112 Core Indicators within each of 56 sub-themes, which identifies the measurable criteria for sustainable performance for the sub-theme, Table 1. Core indicators are applicable at the macro level – meaning to all enterprise sizes and types, and in all contexts. Core indicators serve the purpose of providing standardized metrics to guide future assessments on sustainability. The core indicators’ set is needed for a general level of reporting, as SAFA users do not necessarily have the knowledge to develop indicators themselves, without the risk of lowering the bar of the assessment. The overview of core indicators per sub-themes and themes of Good governance and Environmental integrity are introduced in Table 1 and 2. Core indicators per subthemes and themes of economic resilience and social well-being [18] are not presented, because of extent of the paper. Dimension G: GOOD GOVERNANCE Themes G1 Corporate Ethics G2 Accountability Sub-Themes G1.1 Mission Statement Core Indicators G 1.1.1 Mission explicitness G 1.1.2 Mission driven G 1.2 Due Diligence G 1.2.1 Due diligence G 2.1 Holistic Audits G 2.1.1 Holistic audits G 2.2 Responsibility G 2.2.1 Responsibility G 2.3 Transparency G 2.3.1 Transparency Dimension G: GOOD GOVERNANCE Themes Sub-Themes Core Indicators G 3.1.1 Stakeholder identification G 3.1 Stakeholder Dialogue G3 Participation G4 Rule of Law G 3.1.3 Engagement barriers G 3.1.4 Effective participation G 3.2 Grievance Procedures G 3.2.1 Grievance procedures G 3.3 Conflict Resolution G 3.3.1 Conflict resolution G 4.1 Legitimacy G 4.1.1 Legitimacy G 4.2 Remedy, Restoration and Prevention G 4.2.1 Remedy, restoration and prevention G 4.3 Civic Responsibility G 4.3.1 Civic responsibility G 4.4 Resource Appropriation G5 Holistic Management G 3.1.2 Stakeholder engagement G 4.4.1 Free, prior and informed consent G 4.4.2 Tenure rights G 5.1 Sustainability Management Plan G 5.1.1 Sustainability management plan G 5.2 Full-cost Accounting G 5.2.1 Full-cost accounting Table 1. Overview of core indicators per sub – themes and themes of good governance [18] 3.2 GRI methodology In June 2013, the GRI released the G4 Sustainability Reporting Guidelines [21] updating the G3.1 Guidelines from 2011 [3] and updating and completed the G3 Guidelines from 2006 [4]. The new G4 Sustainability Reporting Guidelines (the Guidelines) [21] offer Reporting Principles, Standard Disclosures and an Implementation Manual for the preparation of sustainability reports by organizations, regardless of their size, sector or location. Dimension E: ENVIRONMENTAL INTEGRITY Themes E1 Atmosphere Sub-Themes E 1.1 Greenhouse Gases (GHG) Core Indicators E 1.1.1 GHG reduction target E 1.1.2 GHG mitigation practices E 1.1.3 GHG balance E 1.2 Air Quality E 1.2.1 Air pollution reduction target E 1.2.2 Air pollution prevention practices E 1.2.3 Ambient concentration of air pollutants E2 Water E 2.1 Water Withdrawal E 2.1.1 Water conservation target E 2.1.2 Water conservation practices E 2.1.3 Ground and surface water withdrawals E 2.2 Water Quality E 2.2.1 Clean water target E 2.2.2 Water pollution prevention practices E 2.2.3 Concentration of water pollutants E 2.2.4 Wastewater quality E3 Land E 3.1.1 Soil- improvement practices E 3.1.2 Soil physical structure E 3.1 Soil Quality E 3.1.3 Soil chemical quality E 3.1.4 Soil biological quality E 3.1.5 Soil organic matter content E 3.2 Land Degradation E 3.2.1 Land conservation and rehabilitation plan E 3.2.2 Land conservation and rehabilitation practices E 3.2.3 Net loss/gain of productive land E4 Biodiversity E 4.1 Ecosystem Diversity E 4.1.1. Landscape/marine habitat conservation plan E 4.1.2 Ecosystemenhancing practices E 4.1.3 Structural diversity of ecosystems E 4.1.4 Ecosystem connectivity E 4.1.5 Land–use and landcover change E 4.2 Species Diversity E 4.2.1 Species conservation target E 4.2.2 Species conservation practices E 4.2.3 Diversity and abundance of key species E 4.2.4 Diversity of production E 4.3 Genetic Diversity E 4.3.1 Wild genetic diversity enhancing practices E 4.3.2 Agro-biodiversity insitu conservation E 4.3.3 Locally adapted varieties/breeds E 4.3.4 Genetic diversity in wild species E 4.3.5 Saving of seeds and breeds E5 Materials and Energy E 5.1 Material Use E 5.1.1 Material consumption practices E 5.1.2 Nutrient balances E 5.1.3 Renewable and recycled materials E 5.1.4 Intensity of material use E 5.2 Energy Use E 5.2.1 Renewable energy use target E 5.2.2 Energy-saving practices E 5.2.3 Energy consumption E 5.2.4 Renewable energies E 5.3 Waste Reduction and Disposal E 5.3.1 Waste reduction target E 5.3.2 Waste reduction practices E 5.3.3 Waste disposal E 5.3.4 Food loss and waste reduction E6 Animal Welfare E 6.1 Health and Freedom from Stress E 6.1.1 Integrated animal health practices E 6.1.2 Humane animal handling practices E 6.1.3 Species-appropriate animal husbandry E 6.1.4 Freedom from stress E 6.1.5 Animal health Table 2. Overview of core indicators per sub – themes and themes of environmental integrity [18] The Guidelines also offer an international reference for all those interested in the disclosure of governance approach and of the environmental, social and economic performance and impacts of organizations. Food Processing Sector Supplement (FPSS) [32] Guideline provided organizations in this sector a tailored version of the G3.1 Guidelines. It includes the original Guidelines, which set out the Reporting Principles, Disclosures on Management Approach and Performance Indicators for economic, environmental and social issues. Sector-specific categories on Sourcing and Animal Welfare are included for the food-processing sector. The Supplement’s additional commentaries and Performance Indicators, developed especially for this sector, capture the issues that matter most for food processors. The G4 Guidelines consists of two parts: 1) Reporting Principles and Standard Disclosures and 2) Implementation Manual. The first part of the Guidelines contains Reporting Principles, Standard Disclosures, and the criteria to be applied by an organization to prepare its sustainability report “in accordance” with the Guidelines. Definitions of key terms are also included. The second part contains explanations of how to apply the Reporting Principles, how to prepare the information to be disclosed, and how to interpret the various concepts in the Guidelines. References to other sources, a glossary and general reporting notes are also included. Standard Disclosures are divided as follows: General Standard Disclosures (Strategy and Analysis; Organizational Profile; Identified Material Aspects and Boundaries; Stakeholder Engagement; Report Profile; Governance and Ethics and Integrity) and Specific Standard Disclosures (Guidance for Disclosures on Management Approach; Guidance for Indicators and Aspect-specific Disclosures on Management Approach). The GRI guidelines are probably the most comprehensive guide covering both what information should be reported and how this information should be provided. 3.3 Integrated reporting methodology Integrated reporting as proposed by the International Integrated Reporting Council (IIRC) [27] is a form of corporate reporting that provides a clear and concise representation of how an organization creates value, now and in the future. The current definition of “integrated reporting” according to the IIRC, is “a process that results in communication, most visibly a periodic ‘integrated report’ about value creation over time. An integrated report is a concise communication about how an organization’s strategy, governance, performance and prospects lead to the creation of value over the short, medium and long term.” Essentially, the goal is to promote “integrated thinking” in a manner that tells the “corporate story” for investors on how financial and nonfinancial reporting line up and how true performance is determined by equating all features of dayto-day operations. The IIRC has proposed the elements to be presented in an integrated report through the Integrated Reporting Framework they are now developing. From 16 April 2013, for the 90 days leading up to 15 July 2013, the IIRC was calling on all stakeholders across the world to read the Consultation Draft of the International <IR> Framework [28]. The IIRC asks stakeholders to understand it, challenge it, critique it, and feedback to us what parts you feel work and what parts perhaps do not. Final version of the Integrated Reporting Framework will be published in the second half of 2013. 3.4 Other methodologies There are a number of alternative reporting methodologies that focus more on what should be reported, rather than how it should be reported. An example is AEI EU (Agri-Environmental Indicators of European union). The defined AEIs EU [1] are not yet generally usable in practical generic reporting for SMEs in the agricultural sector [9]. Some of the indicators cannot be computed because of missing data; sometimes the data required are not homogenous enough [7]. Because these indicators are designed mainly for the purpose of CAP EU (Common Agricultural Policy of EU) they should be considered usable only for special cases of farms. As such, they are divided into completely different domains and subdomains, because the Driving force – State – Response framework is used (responses, driving forces, pressures and benefits and state/impact). In addition, they focus solely on environmental monitoring and are completely lacking both economic and corporate governance domains [8]. The same problems are brought by the application of indicators tracked by the European Environment Agency [13]. Again, these concentrate solely on the environmental domain and are unsuitable for corporate performance reporting. The main reason for the evaluation of these AEIs was that they are highly desirable for the development of ICT tools, which will be capable of producing the required values of these indicators if necessary. Another set of standardized performance indicators comes from the collaboration of OECD, Eurostat, and the Food and Agriculture Organization (FAO) [17]. Although these indicators are not based on the Driving force – State – Response framework, they are still primarily agroenvironmental indicators and are therefore insufficient for corporate sustainability reporting. Within the solution of the project GACR 403, we have proposed a minimal set of generic key performance indicators [10], [14] [19] and [23] applicable to a wide range of enterprises. This forms one set of indicators considered for implementing and validating the information system prototype. The second set of testing indicators are indicators described in Methodology for assessing the sustainability of crop production systems for the conditions of the Czech Republic [9], [16], [20]. The rationale behind these choices is that we want to test both simpler (and generic) reporting systems such as the one proposed by [14] and corporate reporting systems of indicators such as SAFA [7], [18] or SAGROS (Sustainable AGROSystems) [16], [20]. Our goal is to provide an architectural proposal of ICT tools that will be generic enough to contain all of the above-described methodologies. To implement such an information system successfully, several abstractions are required. They need to be carefully constructed to ensure that they are hidden before the end-user and that they do not add more complexity to the corporate reporting itself. 4 System requirements and objectives In our vision, we foresee two main objectives of the corporate reporting information system. A company may use such a system to create various reports and share data with both its stakeholders and, in the future, with the state administrations concerned with regulatory demands and mandatory corporate reporting [29]. The second objective is the corporate performance assessment performed by evaluating key performance indicators by the means of one of the proposed methodologies. This way, the company can share with the public its performance or check a performance development progress. Evaluating corporate sustainability performance can be done through various reports as stated above or through custom dashboards. We define the following three main generic actors: Reporter – a person responsible for ensuring company mandatory reporting (especially financial, social and environmental) or reporting for stakeholders; this would typically be a member of the accounting team or a contractor. This is the typical actor who would enter data into the information system. Evaluator – a person responsible for evaluating company performance; this role can be represented by a wide range of people starting with managers, auditors and various internal interested parties. The evaluator may also enter data into the information system, but it is not his/her primary task, typically this would be the additional data required for the report generation, but not required as the part of mandatory reporting (e.g. company strategy, goals a targets, vision etc.) Administrator – the administrator of the information system responsible for defining the report templates and business rules used to generate reports. The reporting system should include a flexible layout design, rich visualization, business requirements, and data logic definition etc. following Reporting Principles of the G4 Guidelines. In addition, there is the need for translation or publishing requirements, a central deployment and the customization of interactive reports. These primary functions represent a user-friendly reporting system with strong delivery capabilities of relevant reports. From the definition of the generic roles we can proceed to basic functional requirements. Basic high-level requirements for the reporting information system are: store source company data; compute performance indicators according to selected methodology; generate reports in the selected format for offline storage, printing and archiving; import source company data from the main company information system; provide selected information in the form of reports accessible online to the general public or only to selected persons; provide selected information in the standardized XBRL format for interchange with other systems; provide the possibility to evaluate company performance anonymously; provide the possibility to define report data and business logic; dynamically generate the XBRL taxonomies for a given report; provide the possibility to customize standardized reports and to generate fully customized reports; provide the possibility to compare reports. Non-functional requirements for the reporting information system are: the system must be easy to use and accessible to the general public (not just to a few data analysts); as many indicator values as possible should be computed or imported from external sources; the system must be able to provide the software as a service (SAAS) for cloud computing; report layouts must be flexible; report visualizations should be rich and interactive. For all the use cases of the generic actors – reporter and evaluator – we consider two scenarios. The system may be used either on a regular basis (a registered company scenario) to generate scheduled company reports (annual, quarterly, etc.), performance and to check the trailing corporate sustainability performance, and its development. The second scenario (an anonymous company scenario) represents the case when a company wants to generate a one-time irregular report. The second scenario can occur for example when the company is applying a report for a subsidy or grant or when the company is evaluating changes in the company operation (e.g., restructuring), where a report would be generated for the period before the change and for the period after the change. The second scenario can also be seen as an entry point into the reporting system for companies that do not use it regularly. 5 Information System structure The reporting information system consists of several modules that can be separated into two main groups – data collection and data retrieval. The data collection module may occur for the two scenarios described above. Hence, it has two main modules: RegisteredCompanyModule and SurveyModule. The former allows entering data on a regular basis by registered users for companies registered in the information system. The latter allows entering data anonymously, for several years at once. The third data retrieval module is the ImportModule that represents an interface between the information system and its external sources. This can be used in conjunction with both the RegisteredCompanyModule and the SurveyModule. The import module must be able to import data in XBRL format, it should also be able to import data from other sources, but this is highly dependent on the performance assessment methodology used. In these modules are also included the operations of document validation and the application of defined business rules [5], [6]. The data retrieval modules also include the ReviewModule which is a basic document viewer used for reviewing entered data regardless of the reporting framework used. The second important module of data retrieval is the ReportModule that generates the reports according to the selected methodology for the end-user. Both modules may use several different internal modules, which provide the system with indicator mapping, and additional processing of optional business rules regarding the report output (e.g., how often should the report be generated, uniform visual style, etc.). The internal modules include the XBRLReportModule that provides the data in the form of a XBRL document; this module may also be used directly for exporting reports in an interchangeable format. The above-described architecture describes the portion of the information system built as a transactional system utilizing a relational database, which forms the main data storage system. We have proposed a further extension of the reporting information system with XML and XBRL data store to provide greater flexibility [5], [6]. 6 Indicator abstraction To encompass the above-defined rules and different methodologies described in the previous chapter, we propose the following abstraction of performance indicators (applicable universally to performance indicators [23]). For the prototype software application, we have used a relational database as a storage system. Hence, the following description uses relational database technology. See the chapter 8 for the discussion of other storage options. The basic indicator abstraction entities are: indicator group; indicator; indicator item; indicator item aspect. Fig. 2. Structuring of indicators with sample concrete indicators from methodology [16] These entities are organized into a hierarchical structure shown on Fig. 2. Each indicator sub-entity is shown along with examples of real values according to the methodology SAGROS [16]. The end user will always fill in the leaves of this tree-structure – i.e., if an indicator value is not computed and contains only one value, it may contain only a single indicator item, which will contain no aspects. The concrete example may be an indicator average number of employees, which is the value directly known to the company and a value directly used in a report. It is important to note that an indicator group must always have at least one indicator and all the indicators must have at least one indicator item while an indicator item may have zero or more indicator item aspects. This is important in order to maintain the flexibility of the reporting information system. In case the indicator item contains several aspects, it is expected that all of them will be required for the indicator item value computation, unless specified otherwise in the indicator item formula. The entities described above define the core of the data model of the information system. The following schema (0) shows further details with the associated entities. Here we have defined a report type, which represents an instance of a given methodology, e.g., the SAFA version 2.0 [18]. We have also introduced a report instance that represents a concrete filled-in report for a specific company for a specified time period. The indicator value, which is an actual value entered by the end-user, may be bound to either indicator item or indicator item aspect, but not both. Note that the indicator value is not actually related to a report instance. This is internal so that a single indicator value may be reused in several report instances. For example, the value of the indicator average number of employees may be used both in the SAFA 2.0 report for the year 2012 and also for the mandatory financial report for the same period. As long as the period is the same, the value should be reusable. This is crucial to maintaining the consistency of data across different reports and it is a highly important feature in order to make reporting as easy and quick as possible. There is also the indicator report group entity, which defines grouping of indicators in the defined report types. The linkage is provided by the indicator report-type association table, which assigns indicators to specific groups in specified report types. Note that it is possible that an indicator is used in more than one report definition, which is desired behaviour. It is also possible that an indicator is used in more than one group, which is also desired behaviour. For example, the indicator average salary may be part of the economic indicators group in the report, or it may be part of the social responsibility group, or even both. The actual computing of the reported indicator value is provided by the formula code defined in the indicator, he indicator item, and the indicator item aspect entities. The formula API (Application Programming Interface) is discussed further in the text. The computed value is not stored in the schema shown in 0, because it is recoverable from the base values. However, for performance and/or version tracking reasons, the computed values may be stored in the caching schema. The entities: indicator value, indicator item, and indicator item aspect all link to data type. The value data types model is shown in 0. Here we define another entity named Core data type, which represents an XBRL native type (e.g., monetary, numeric, fractional, etc.). This is because the XBRL data types are too coarse for all kinds of reports. The definition of a constraint also uses the formula API in the following chapter. Fig. 3: Core data model of the reporting system Closely linked to an indicator value computation is also the Associated Indicator table. This is used to track dependencies between indicators in a case when one indicator is used for the computation of another indicator. The concrete example can be seen in the methodology [15] where the economic performance indicator EE14 – Expenses on R&D is computed as the value of the Total costs of R&D of a company divided by the value of the economic indicator EE01 – Value added of the company and expressed in per cent. The EE01 – Value added is the economic performance indicator of [14] defined as the value added of the company per employee. The further economic performance indicator EE02 – Value added per personal costs was introduced [14]. Both EE01 and EE02 indicators are computed as the value added of the company divided by the average number of its employees in the year for the former, and by the payroll costs, for the latter. To fully represent the above-mentioned indicators, we need to define the following indicators in the application: source indicator: total costs of R&D; source indicator: average number of employees; source indicator: payroll costs; source indicator: value added of the company; output indicator: EE14 – Expenses on R&D; output indicator: EE01 – added value per employee; output indicator: EE02 – added value to personnel costs. Note that the separation of indicator groups into source and output indicators is purely explanatory, there is no such division made in the data model as any indicator may be used in any report. Because of this, it is important to maintain associations between indicators (or potential indicators) so that when the user decides to fill a given type of report, all of the required values are requested – that is, all the Core indicators, see 0. Fig. 4: Data type data model 7 Indicator computation The actual indicator values for a report are computed from the indicator items and indicator item aspects. The computing itself is provided by a formula defined in each of these entities. The formula is a piece of executable code used to compute the value from the source data. In our prototype of the reporting information system, the formula is currently defined as a PHP code. This is the native language of the information system. In the future it might be better to switch to another interpreted language such as JavaScript, which would provide more flexibility and portability. Regardless of the formula language used, the following data API is proposed. Each indicator may have associated indicator values; these are the values used in more indicators at once. The concrete example would be an indicator added value, which would be used as both a report indicator and also as a common divisor for several other indicators in the methodology [15]. The API for the formula code has one available base entity: indicator. The properties defined in each entity correspond to those defined in the data model. The computation must return a single value of the arbitrary data type. We expect that a single report indicator must always have a single value. This may pose a limitation to the report generator, because the end user might want to obtain a list of values in the report. Currently, it is only possible to convert the data type among the indicator, indicator item and indicator item aspect; it is also possible to generate strings of characters from raw input values. In the future development we will see whether such a solution is sufficient or whether multivalued data types will be required. When working with the formulas, the question of versioning comes into attention. All methodologies are being constantly updated and the computations may change, so it is required to keep track of historical data – the definitions of indicators. We have proposed a standard recording of history, where after each change in the formula a new physical indicator is created, while maintaining the link to the original indicator. That is, when an indicator is modified, a new indicator is created, one that supersedes the original indicator. The above principle applies to all core entities of the data model (i.e., indicators, indicator items, indicator item aspects). 8 Other storage option concepts Current trends in reporting are causing that the usage of the reporting systems is moving from thee specialized usage by a few data analysts to general and ubiquitous use across many organizational units and users. In addition, data storages with simply historical information are moving to real-time analysis and predictive analysis of the future. When we focus more on the organization unit, we can see how formerly fragmented views with silos of information – in many cases stored in transactional systems – are being brought together to enterprise wide and unified views. There are two basic approaches a company can take when trying to achieve such unified views – either a centralized data warehouse is built, or a series of connectors is built between various systems. Companies are also searching for time-relevant and critical data using pre-built Business Intelligence (BI) applications [5], [6], [31]. This is being required more than just individual reports based on as-needed basis built within dedicated analytics tools. When comparing classic reporting tools (also called traditional reporting tools) against an approach based on XBRL, we expect that the classic reporting tools will have slower development cycles, higher cost of development and maintenance. The main reason for this being the fact that prebuild tools (e.g. Oracle Hyperion) may be used to transform XBRL data directly into reports without the need of programing a new application. In our previous research [5], we already mentioned many XBRL and inline XBRL (iXBRL) technical challenges. Fig. 5: Overview of system architecture built on XBRL database For example, we discussed how to improve core processing engine and integrity with business rules approach in [5], or how to connect BI with XBRL in [6]. There are still more challenges which need to be resolved such as an efficient queryability and analytics and scalable lookups over large volumes of XBRL documents. As we mentioned in [6], XBRL was not explicitly designed for the role of a core storage system; it was designed as a metadata representation language. However, it is possible for XBRL to become a native data source in the areas of text mining and web reporting. With this approach, we can easily eliminate the most of the costly process of data preparation. The reason why the proposed reporting information system should work is that existing software or other mapping tools can do transforming of XBRL data automatically. This automation can increase the consistency, accuracy and trust in the data – all key tenets of successful reporting. 0 shows an overview of a system architecture based on an XBRL database. The top application layer accepts report submissions and provides results of report queries. It can be composed of various interfaces such as web application user interface or some defined APIs. The application layer communicates with the storage layer via SQL queries or via XQuery language. The data layer is composed of both the RDB system and the XBRL database. The data layer exposes both relational and XML interface so that the data can be queried by both SQL and XQuery languages. The application layer may therefore provide a wide range of APIs to connect the database to various sources. This offers a high interoperability of the system, which is a necessary feature. Aside to the application layer is the business intelligence engine, which can use direct access to the data layer. This will be possible if the data layer is self-containing – i.e. it can provide a XBRL document with all necessary taxonomies and templates. Our proposed architecture allows easy integration of XBRL processing with third party business intelligence software products to satisfy diverse use cases [31]. This solution allows the XML-based querying and protocol access. On 0, we describe on how a range of APIs and Internet protocols may be used to access the source data. For example, SOAP, REST, JDBC or file based access to the content can be provided with WebDAV interface. Important feature of this approach is the storage of XBRL in its original XML representation while still providing relational access. The background is a relational representational view that is built over the XBRL content by exposing set of base entities physically implemented as relational views over XBRL documents. RDBMS logical data model and XBRL content together provide an easy integration with BI tools. Direct access to source data can be used to perform a wide variety of reporting and charting outputs. Native XBRL storage with the integrity based on XBRL rules allows both XML and SQL querying based on XBRL semantics. The main feature is the connection with the current transactional database via relational application integration. This is an important feature for those enterprises, which do not have data in the form of XBRL documents. There is a clear need for case studies and research pilots that would test online analytics based on XBRL dimension (explicit, typed). 9 Conclusion We have implemented a prototype of the described special reporting information system. There have been two primary objectives of the prototype implementation: to verify that the proposed indicator abstraction can be implemented successfully and to verify the usability of the reporting information system for end-users. Though our main objective is the implementation of the SAFA Guidelines methodology [18], we have currently implemented the Methodology for assessing the sustainability of crop production systems for the conditions of the Czech Republic [16], because the SAFA Guidelines version 2.0 have been published only recently. As far as usability is concerned, we are currently testing two approaches to user interface – survey interface, and guide interface. In the former interface, the user can enter all the required and optional values on a single page form with minimum structuring. In the latter case, the user is lead through various steps, filling different parts of the report gradually. Following the SAFA vision [18], we use four dimensions of sustainability: good governance, environmental integrity, economic resilience and social well-being, 21 universal Themes and 56 Sub-themes including 112 Core indicators. The proposed reporting information system is able to deal with other corporate sustainability performance assessment frameworks and guidelines. Our future work will concentrate on improving the XBRL reporting by adding dynamic XBRL taxonomies and an XBRL database layer to the GRI XBRL taxonomy proposal. We will also concentrate on improving the end-user experience of the report outputs so that performance evaluation becomes available to a wide range of users. This will be accomplished by utilizing business intelligence tools to add a set of dashboards and overviews to the report [5]. 10 Acknowledgment This paper is supported by the Czech Science Foundation. Name of the Project: Construction of Methods for Multifactor Assessment of Company Complex Performance in Selected Sectors. Reg. No P403/11/2085. 11 References [1] Analytical framework. European Commision. Eurostat Home [online]. 2010, Avaliable from : http://epp.eurostat.ec.europa.eu/portal/page/portal/agri_environmental_indicators/introd uction/analytical_framework [2] Z. Chvátalová, A. Kocmanová and M. Dočekalová, “Corporate Sustainability Reporting and Measuring Corporate Performance,” in Proceedings of 9th IFIP WG 5.11 International Symposium. ISESS 2011. Environmental Software Systems. Frameworks of eEnvironment. J. Hřebíček, G. Schimak, R. Denzer, Eds. Heidelberg: Springer, 2011, pp. 398–406. [3] G3.1 Guidelines on Globalreporting. 2011. [Online]. https://www.globalreporting.org/reporting/G3andG3-1/g3-1guidelines/Pages/default.aspx Available from: [4] G3 Guidelines website on Globalreporting. 2006. [Online]. Available from: https://www.globalreporting.org/reporting/G3andG3-1/g3-guidelines/Pages/default.aspx [5] M. Hodinka and M. Štencl, “Reporting Tools for Business Rules,” In IT for Practice 2011. Ostrava: VŠB Technická univerzita Ostrava, 2011, pp. 29-36. [6] M. Hodinka, M. Štencl, J. Hřebíček and O. Trenz, “Business intelligence in Environmental reporting powered by XBRL,” Acta univ. Agric. Et silvic. Mendel. Brun., 2013. in press [7] J. Hřebíček, O. Trenz and E. Vernerová “Optimal set of agri-environmental indicators for the agricultural sector of Czech Republic,” Acta univ. agric. et silvic. Mendel. Brun., 2013. in press. [8] J. Hřebíček, M. Popelka, M. Štencl and O. Trenz, “Corporate performance indicators for agriculture and food processing sector,” Acta univ. Agric. Et silvic. Mendel. Brun., Vol. 60, pp. 99-108, February 2012. [9] J. Hřebíček, S. Valtyniová, J. Křen, M. Hodinka, O. Trenz and P. Marada, “Sustainability indicators: development and application for the agriculture sector,” in Sustainability Appraisal: Quantitative Methods and Mathematical Techniques for Environmental Performance Evaluation, M. G. Erechtchoukova, P. Khaiter and P. Golinska, Eds. Heidelberg: Springer, 2013, pp. 63-102. [10] J. Hřebíček, M. Štencl, O. Trenz and J. Soukopová, “Current Trends in Corporate Performance Evaluation and Reporting in the Czech Republic,” International Journal of Energy and Environment, WSEAS Press, Vol. 6, pp. 39-48, January 2012. [11] J. Hřebíček, M. Hodinka, M. Štencl, O. Popelka and O. Trenz. “Sustainability Indicators Evaluation and Reporting: Case Study for Building and Construction Sector“. In N. Altawell, K. Volkov, C. Matos, P. F. de Arroyabe. Recent Researches in Environmental and Geological Sciences. Proceedings of the 7th WSEAS International Conference on Energy & Environment (EE '12). USA: WSEAS Press, 2012. pp. 305-312, [12] ISEAL Alliance. Assessing the impacts of environmental and social standards systems. ISEAL Code of Good Practice, version 1.0. ISEAL Alliance, London Available from: http://www.isealalliance.org/resources/p041-impactscode-of-good-practice. [13] Indicators: European Environment Agency. European Environment Agency [online]. Copenhagen, 2013. Available from: http://www.eea.europa.eu/data-and-maps/indicators [14] A. Kocmanová and M. Dočekalová, “Construction of the economic indicators of performance in relation to environmental, social and corporate governance (ESG) factors,” Acta univ. agric. et silvic. Mendel. Brun., Vol. 60, pp. 195–205. April 2012. [15] A. Kocmanová, M. Dočekalová, P. Němeček, and I. Šimberová, “Sustainability: Environmental, Social and Corporate Governance Performance in Czech SMEs,”. in The 15th World Multi-Conference on Systemics, Cybernetics and Informatics. IFSR, Orlando, USA: WMSCI, 2011, pp. 94–99. [16] J. Křen, Methodology for assessing the sustainability of crop production systems for the conditions of the Czech Republic. Brno: Mendel University in Brno, 2011. [17] OECD. Environmental Indicators for Agriculture Volume 1: Concepts and framework [online]. Paris: OECD, 1999 Avaliable from: http://www.oecd.org/agriculture/sustainableagriculture/40680795.pdf [18] SAFA Sustainability Assessment of Food and Agriculture systems. Draft Guidelines (Version 2.0), 2013, Available from: http://www.fao.org/fileadmin/templates/nr/sustainability_pathways/docs/SAFA_Guidelin es_final_draft.pdf. [19] J. Soukopová and E. Bakoš, “Assessing the efficiency of municipal expenditures regarding environmental protection,” in Environmental Economics and Investment Assessment. III. ed. Cyprus. WIT Press. 2010, pp. 107 –119. [20] S. Valtýniová and J. Křen,”Indicators used for assessment of the ecological dimension of sustainable arable farming – review,”, Acta univ. agric. et silvic. Mendel. Brun., Vol. 59, pp. 247–256, March 2011. [21] G4 Guidelines website on Global reporting. 2013. [Online]. Available from: https://www.globalreporting.org/reporting/g4/Pages/default.aspx [22] O. Popelka, M. Hodinka, J. Hřebíček and O. Trenz “Information System for Corporate Performance Assessment and Reporting”. In: In Charles A. Long, Nikos E. Mastorakis, Valeri Mladenov. RECENT ADVANCES in SYSTEMS, CONTROL, SIGNAL PROCESSING and INFORMATICS. Proceedings of the 2013 International Conference on Systems, Control, Signal Processing and Informatics (SCSI 2013). EUROPMENT, 2013. pp. 247-253 [23] Z. Chvátalová and I. Šimberová, “Analysis and identification of joint performance measurement indicators: Social and corporate governance issues”. Acta univ. agric. et silvic. Mendel. Brun., Vol. 60, pp. 127-138, July 2012. [24] L. Pinter, P. Hardi, A. Martinuzzi and J. Hall, ”Bellaggio STAMP: Principles for sustainability assessment and measurement”. Ecological Indicators Vol. 17, pp. 20-28, 2011. [25] ISEAL Alliance. Assessing the impacts of environmental and social standards systems. the ISEAL Credibility Principles (ISEAL Credibility Principles v0 3 - 11 June 2013. ISEAL Alliance, London Available from:http://www.isealalliance.org/resources/p041impactscode-of-good-practice. [26] GSCP. 2010. Reference Tools of the Global Social Compliance Programme. Issy-lesMoulineaux, France Available from http://www.gscpnet.com/working-plan.html. [27] IIRC. 2013. [Online]. Available from: http://www.theiirc.org/the-iirc/. 2013. [28] Consultation Draft of the International <IR> Framework [Online]. Available from: http://www.theiirc.org/wp-content/uploads/Consultation-Draft/Consultation-Draft-of-theInternationalIRFramework.pdf [29] J. Hřebíček, F. Piliar and J. Soukopová. “Case study of voluntary reporting in SMEs“. In: Sustainability Accounting and Reporting at macroeconomic and microeconomic level. pp. 265-273, Linde, Praha, 2009. [30] J. Hřebíček, J. Soukopová and E. Kutová. “Standardisation of Key Performance Indicators for Environmental Management and Reporting in the Czech Republic“. In USCUDAR 2010. Recent Advantages in Urban Planning, Cultural Sustainability and Green Development. USA: WSEAS Press, 2010. pp. 76-82. [31] D. Airinei and D. Homocianu. “Data Visualization in Business Intelligence“. In RECENT ADVANCES in MATHEMATICS and COMPUTERS in BUSINESS, ECONOMICS, BIOLOGY & CHEMISTRY. Proceedings of the 11th WSEAS International Conference on MATHEMATICS AND COMPUTERS IN BUSINESS AND ECONOMICS (MCBE '10). USA: WSEAS Press, 2010. pp. 164-167. [32] Food Processing Sector Supplement, 2012. [Online]. Available from: https://www.globalreporting.org/reporting/sectorguidance/sector-guidance/foodprocessing/Pages/default.aspx Generování japonsko-českého slovníku Jonáš Ševčík Masarykova univerzita - Fakulta informatiky Laboratoř softwarových architektur a informačních systémů Botanická 68a, 602 00 Brno 255493@mail.muni.cz Abstract. This Article aims to present the reader with methods suitable for automatic bilingual dictionary generation. Proposed methodology uses automatic generation using pivot-language method. Input data are Japanese-English and English-Czech dictionaries, where using WordNet and pivot-language method we generate Japanese-Czech corpus, which will be used in online collaborative service suitable for finishing the rest of sense translation. Abstrakt Tento článek si klade za cíl seznámit čtenáře s principy automatizovaného generování dvojjazyčného slovníku. Navržená metodologie uplatňuje automatické generování pivotovou metodou. Jako vstupní data jsou navrženy japonsko-anglický a anglicko-český slovník, kde za pomocí WordNetu a pivotové metody je získán japonsko-český korpus, který následně bude použit v online kolaborační službě určené pro crowdsourcingové dotvoření zbylého překladu. Keywords Bilingual dictionary, automatic generation, Japanese, Czech, crowdsourcing Klíčová slova Dvojjazyčný slovník, automatizované generování, japonština, čeština, crowdsourcing 1 Úvod České prostředí postrádá elektronický japonsko-český slovník, který by alespoň částečně dosahoval kvalit slovníku JMDict [1], případně jeho lokalizovaným verzím. Samozřejmě zde existují papírové verze slovníků, jako je např. Japonsko český slovník [3] nebo Japonsko český studijní znakový slovník [4]. V současné době však neexistuje ucelený japonsko-český slovník v elektronické podobě. Dostupnost takovéhoto slovníku je nezbytná nejen pro studium Japonštiny. Užití překladů do jiného než rodného jazyka uživatele vyžaduje netriviální znalost cílového jazyka, kde navíc bez uvedení kontextu může dojít k chybné interpretaci jednotlivých hesel. Tvorba slovníku není triviální a vyžaduje kvalifikované lidské zdroje. Jedná se tedy o nákladnou činnost, která je v případě tvorby jednotlivcem též časově náročná. Je proto potřeba navrhnout řešení, které by umožnilo tvorbu slovníku realizovat levněji a s co nejmenšími nároky na lidské zdroje. Tento výzkum si dává za cíl navržení univerzálně použitelného nástroje pro kolaborativní tvorbu elektronického japonského slovníku. Data by tedy měla být tvořena samotnými uživateli slovníku. Pro vytvoření základního korpusu je zamýšleno použití automatického generování obsahu. Při automatizované tvorbě obsahu se nabízí řešení použití volně dostupného slovníku, např. zmiňovaného Edict [1] a jednotlivá hesla přeložit doslova spárováním s jiným např. anglickočeským slovníkem. Toto řešení překladu přes společný jazyk je však naivní. Jednotlivá hesla slovníku můžou obsahovat významové posuny, případně bez uvedení příkladových vět může dojít k jejich chybné interpretaci. 1.1 State of the art Během 80. let 20. Století byla vyvinuta řada metodik pro automatické generování dvojjazyčných slovníků. Za zmínku stojí Dice-coefficient, Corespondence-tables, či Mutual information [4]. Nevýhodou těchto přístupů je, že vyžadují rozsáhlé dvojjazyčné korpusy. V roce 1994 představil Tanaka [5] nový přístup založený na tzv. pivotovém jazyku. Jako zdroj jsou použity slovníky s překlady do a z pivotového jazyka. Tato metoda spočívá ve vyhledávání slov přes třetí zprostředkovatelský jazyk. A takovýto přístup samozřejmě může být automatizován. Metoda používá pro automatizovaný překlad dvou obousměrných slovníků2. Správnost překladu je ověřována přeložením hesla v jednom a následně zpětném směru. Podobný přístup navrhl také Sjöbergh [6], který pro tvorbu japonsko-švédského slovníku využil angličtiny jako pivotovacího jazyka. Rozšíření uvedené pivotové metodiky existuje celá řada. Ovšem nejedná se o obecně použitelné techniky. Vychází především ze specifik použité kombinace jazyků nebo jejich podobnosti. Např. Širai [7] využívá generování japonsko-korejského slovníku přes angličtinu, ovšem využívá též specifik dřívějšího zápisu korejštiny pomocí hanzi (čínských znaků) a pro překlad sino-korejských slov využívá přímé mapování na slova sino-japonská. V tomto případě je možné najít jednoznačnou rovnost mezi jednotlivými slovy. Podobného mapování je za pomocí zápisu hanzi možno využít i u ostatních většiny substantiv. Je však zřejmé, že s výjimkou jazyků využívajících čínských zanků, případně přejatých čínských slov není možné tyto doplňkové metody použít. Jako další z možných způsobů automatického generování slovníkových dat je využití lingvistické databáze WordNet [8], která definuje pevně danou strukturu pro indexování slov jednotlivých světových jazyků. Za použití indexu je možné mezi sebou jazyky propojovat a vytvořit tak jednoduše dvojjazyčný slovník. Předpokladem je však, že pro daný jazyk je vytvořen WordNet v dostatečném rozsahu. Český WordNet obsahuje pouhých 40000 synsetů3 a japonský přibližně 57000, oproti tomu anglický WordNet obsahuje 117000 synsetů [9, 10, 11]. Je tedy zřejmé, že generované slovníkové data jsou omezeny průnikem těchto porovnávaných struktur. Jako nejvhodnější pro naši potřebu se jeví použití metody publikované Vargou [12]. Ta využívá právě WordNetu a pivotové metody. Ta funguje následovně. Nejdříve se provádí fáze, ve které jsou tranzitivně propojeny jednotlivé slovníky. V našem případě všechna japonská hesla, která mají ekvivalent v angličtině a následně ekvivalent v češtině. Takovéto spojení je nazváno kandidátem pro překlad. Ve druhé fázi proběhne roztřízení kandidátů do 6 skupin, ze kterých jsou vyřazeni ti kandidáti, kteří nesplní požadovanou mez [13]. Výsledky výzkumu [12, 13] ukazují, že z aktuálně dostupných dat, je tímto způsobem možné vygenerovat přibližně 70 % synsetů z celkového průniku propojených WordNetů. 2 Navržené řešení Pro tvorbu japonsko-českého slovníku jsme zvolili pivotovou metodu, jak ji uvádí Varga [12]. Zjevná přednost použití předgenerovaných dat je ta, že je možné tímto způsobem automaticky generovat podstatnou část slovníku. Tento korpus může sloužit jako základ, který bude následně rozšiřován. Jak již bylo zmíněno v úvodu – překlad za pomocí jednotlivce profesionála je časově neefektivní a též finančně náročný. Z toho důvodu jsme se rozhodli vytvořit online službu, která jako svůj základ bude používat japonsko-českého korpusu vygenerovaného výše uvedenou metodou. Pro generování dat bude použito volně dostupného japonského a českého WordNetu. Japonský slovník JMDict (zdroj japonského WordNetu) bude též sloužit jako základ pro české překlady, jelikož obsahuje překlad hesel nejen do angličtiny, ale i němčiny, ruštiny, holandštiny a dalších jazyků. Pro umožnění crowdsourcingového překladu zbývajících nepřeložených hesel to má tedy tu výhodu, že překladatel může vidět u 2 3 Pro češtinu a angličtinu by to byl anglicko-český a česko-anglický slovník. Množin synonym. daného hesla překlady v jiných jazycích a následně doplní český překlad. Využití slovníku JMDict jako základ navíc limituje seznam japonských hesel, takže nebude možné přidávat zcela nová. To má tu výhodu, že nebude umožněno zanést novým japonským heslem chybu, případně neexistující slovo, což by se v opačném případě muselo ověřovat. Všechna data obsažená ve slovníku JMDict již prošla kontrolou, proto je bezpečné je použít a dále na nich stavět. Tento slovník je navíc neustále aktualizován, takže nehrozí jeho případné zastarání a je možné iterativně doplňovat nové přírůstky i do japonsko-českého slovníku. 2.1 Bezpečnostní opatření Navržená služba bude obsahovat strohý uživatelský interfejs. Uživatelům se bude profilovat jako japonsko-český slovník a k dispozici bude pole pro vstupní text a následně omezující kritéria pro hledání řetězce – přímé hledání, hledání prefixu, hledání sufixu a hledání středu zadaného slova. Výsledky budou zobrazovány v českém jazyce, v případě nedostupnosti českého překladu bude uživateli nabídnut cizojazyčný překlad, dle osobních preferencí nadále uložených pomocí cookies. V případě cizojazyčného překladu bude ještě nabídnut strojový překlad do českého jazyka4. Libovolný uživatel pak bude moci navrhnout vlastní překlad daného hesla, případně potvrdit správnost nabízeného strojového překladu. Aby nebyli uživatelé odrazováni od navrhování překladů povinnou registrací, tak nebude použita. Překlady by tak byl schopen navrhnout kdokoli. Aby bylo možné zaručit kvalitu dat, tak čerstvě přeložené hesla nebudou přímo propagována do dočasného slovníku. K tomu, aby bylo možné heslo přenést do slovníku, je potřeba překlad potvrdit. Potvrzení je možné pomocí speciální role administrátora, typicky člověka se znalostí jazyka, který jen rozhodne o správnosti překladu. Není tedy potřeba, aby jej sám překládal. 2.2 Nasazení Pro nasazení služby jsme zvolili cloudové řešení společnosti Red Hat – OpenShift5. Tato platforma nabízí zdarma 3 konfigurovatelné serverové instance, je proto možné si nastavit např. optimální velikost operační paměti. Celá služba je řešena v programovacím jazyce Java, data jsou ukládána do MySQL databáze. Aktuálně se jedná teprve o prototyp služby, proto není veškerá funkcionalita implementována. Nicméně na vývoji se neustále pracuje. 3 Závěr V článku jsme představili možné řešení automatického generování japonsko-českého slovníku. K němu byla zvolena pivotová metodika, která využívá angličtiny jako pivotovacího jazyka. Jednotlivá hesla jsou navzájem provázána pomocí databáze WordNet. Vzniklý japonsko-český korpus bude nasazen v online službě, která nabídne funkcionalitu online s lovníku, s tím, že nevygenerované překlady pro hesla budou strojově doslovně přeloženy a nabídnuty překlady v jiných cizích jazycích. Uživatelé též budou mít možnost poskytnout vlastní překlad k jednotlivým heslům, který bude uložen a nabídnut ke schválení skupinou administrátorů, aby bylo zamezeno výskytu chyb. Jako možné vylepšení služby vidíme možnost eliminovat nutnost existenci administrátorů. Na místo ní by bylo umožněno rozhodovat o správnosti překladu uživatelům s určitým množstvím přeložených a uznaných hesel. A o samotném schvalování výrazů by se rozhodovalo hlasováním, tzn. pokud by navržený překlad obdržel požadovaný počet hlasů, byl by označen automaticky za správný. 4 5 Např. pomocí služby Google Translate. Testovací služba je dostupná na: https://jmd-hotovy.rhcloud.com/Dict/ 4 Literatura [1] JMDict. [online]. [cit. 2013-10-31]. http://www.csse.monash.edu.au/~jwb/wwwjdicinf.html Dostupné z: [2] LABUS, David a Jan SÝKORA. Japonsko český studijní znakový slovník. Praha: Paseka, 2005, 625 s. ISBN 80-718-5736-X. [3] KROUSKÝ, Ivan a František ŠILAR. Japonsko český slovník. Voznice: LEDA, 2005, 519 s. ISBN 80-733-5045-9. [4] BROWN, R.D. 1997. Automated Dictionary Extraction for Knowledge-Free ExampleBased Translation, Proceedings of the 7th International Conference on Theoretical and Methodological Issues in Machine Translation, pp. 111-118. [5] TANAKA, K., UMEMURA, K. 1994. Construction of a bilingual dictionary intermediated by a third language, Proceedings of COLING-94, pp. 297-303. [6] SJÖBERGH, J. 2005. Creating a free Japanese-English lexicon, Proceedings of PACLING, pp. 296-300. [7] ŠIRAI, S., JAMAMOTO, K. 2001. Linking English words in two bilingual dictionaries to generate another pair dictionary, ICCPOL-2001, pp. 174-179. [8] MILLER G.A., BECKWITH R., FELLBAUM C., GROSS D., MILLER K.J. (1990). Introduction to WordNet: An Online Lexical Database, Int J Lexicography 3(4), pp. 235244 [9] BLAHUŠ, Marek. Extending Czech WordNet Using a Bilingual Dictionary. Brno, 2011. Diplomová práce. Masarykova univerzita. [10] Japanese WordNet. [online]. [cit. 2013-10-31]. Dostupné z: http://nlpwww.nict.go.jp/wnja/index.en.html [11] WordNet. [online]. [cit. 2013-10-31]. Dostupné z: http://wordnet.princeton.edu/ [12] VARGA I, JOKOJAMA Š.: Bilingual dictionary generation for low-resourced language pairs. EMNLP '09 Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing. 862-870. [13] VARGA I, JOKOJAMA Š.: iChi: a bilingual dictionary generating tool. Annual Meeting of the Association of Computational Linguistics. 2009. Availability and scalability of web services hosted in Windows Azure David Gešvindr Masarykova univerzita - Fakulta informatiky Laboratoř softwarových architektur a informačních systémů Botanická 68a, 602 00 Brno 255653@mail.muni.cz Abstract This article aims to present the reader with principles of a software architecture design that leads toward better scalability and availability of a web service hosted in Windows Azure cloud environment. This article introduces the reader to major Windows Azure services and points out the most important differences of the cloud environment, that need to be taken in account in an early software architecture design phase so the service can effectively use advantages of the cloud especially almost unlimited horizontal scalability. Abstrakt Tento článek si klade za cíl seznámit čtenáře s principy návrhu softwarové architektury webové služby, jenž vedou k lepším charakteristikám škálovatelnosti a dostupnosti v cloudovém prostředí Windows Azure. Článek seznamuje čtenáře s hlavními službami Windows Azure a upozorňuje na důležité odlišnosti cloudového prostředí, které musí být brány v potaz již při návrhu architektury služby, pokud má služba efektivně využívat výhod cloudu, zejména téměř neomezené horizontální škálovatelnosti. Keywords Cloud computing, Windows Azure, software architecture, scalability, availability Klíčová slova Cloud computing, Windows Azure, softwarová architektura, škálovatelnost, dostupnost 1 Introduction With increasing dependency of modern society on daily used web services even in emergency situations, there are more strict requirements on availability and scalability of hosted services. One of the possible solutions to increase these service characteristics is to use a cloud-based hosting with guaranteed service availability and scalability. Together with broader use of the cloud there are emerging questions if current software architectures used for 2 or 3-tier web applications are able to use full potential of the cloud and if they are also highly available and scalable or there is need to change the architecture of the designed service to ensure this characteristics. In first section of this paper Windows Azure cloud platform is introduced. Section 2 describes recommendations for software architecture to increase scalability of a service. Section 3 describes architecture of a service designed for efficient use of cloud environment. Section 4 evaluates this approach using load tests. 2 Overview of the Windows Azure Platform Windows Azure is Microsoft’s cloud platform providing rich set of hosting services in Microsoft’s datacenters all over the world [1]. It offers hosting services that fall mostly in categories - Infrastructure as a service (IaaS) and Platform as a Service (PaaS). Developers can buy this platform’s compute and storage resources to host their own web applications and services. Considering the size of datacenters of Windows Azure, there is almost unlimited amount of resources that can be provisioned. Payment is based on Pay-As-You-Go model, where the customer is billed only for currently allocated amount of compute power, storage capacity, amount of outgoing traffic from the datacenter, database size, or another billable service metric based on the pricelist of Windows Azure. In the following sections we shortly introduce important services offered by Windows Azure and discuss their key qualitative attributes in the area of availability and scalability. 2.1 Compute Services Windows Azure provides a range of computation services allowing developers to run their applications in the cloud environment. Biggest difference is in a level of control the developer has over the environment – Virtual Machines service provides freedom to build a custom server in contrast to the Web Sites service that offers almost a classical web hosting with minimum configuration options. Developers can create their own virtual server using Virtual Machine service and vertically scale this server by choosing server configuration – number of CPU cores and RAM size. During the creation process they can choose from different available operating system images or they can upload their own hard drive file to Azure Storage and use it in the virtual machine. This service offers high degree of freedom, but also puts a responsibility for server maintenance at the customer. Typical use of this service is a hosting of a specific application server or strongly customized web server. To provide high availability according to the SLA – 99.95%, there have to be at least two instances of the virtual machine grouped together into an availability set. Cloud Service offers lower degree of configuration freedom but also significantly lowers a required maintenance effort. In background it works almost the same way as a virtual machine. Developers can choose between 2 roles – web role and worker role, depending whether they need to host the application in Microsoft’s web server Internet Information Service or run it as a stand-alone process. Application they build is during publication packaged into an installation package together with its configuration and is uploaded to Windows Azure. Windows Azure infrastructure then allocates required computation resources (Fabric Controller creates virtual server from unified server image depending on a selected role, deploys published application to a freshly created server and configures network infrastructure to distribute load between instances of the service if there is more than one). Newly created instance of virtual server can be customized by a startup script, which is a part of the installation package and can reconfigure OS and install another applications. The main disadvantage of cloud service is absence of persistent storage, because the cloud infrastructure never migrates existing virtual server, it always creates new instance based on the application package. Therefore all application data should be stored in Azure Storage or relational database. Cloud Service can scale vertically – increase computing power of a single node or horizontally by increasing a number of nodes and distributing load between them. For web services that do not require any specific configuration changes in a web server it may be cost efficient to use the Azure Web Sites service. This service offers support for ASP.NET, PHP, Node.JS and Python. Web sites can be deployed directly from source controls with support for Git, Bitbucket, CodePlex and Team Foundation Service or via FTP. Their content is stored in the Windows Azure Storage and is loaded directly by web servers, which can run in shared or reserved mode. In a shared mode websites are running on a virtual machine that is concurrently hosting many other websites of different customers in a slightly modified version of Internet Information Service. User in this mode pays for every hosted site. To guarantee better performance it is possible to create instance that is reserved only for one customer and can host up to 500 web sites. Reserved mode is billed in the same way as a virtual machine independently on the number of hosted web sites. Web sites in a reserved mode can be scaled vertically by selecting small, medium or large virtual machine instance and also horizontally by increasing the number of web servers serving the same content. 2.2 Scalability and availability of compute services Benefit of the cloud is that a compute power can be easily increased when it is needed and decreased when the application is idle or under lower load. There are two approaches how a service can be scaled – vertically and horizontally. [2] The vertical scaling increases available resources of a single computation node – it adds more CPU cores and RAM. This is done by changing a configuration of the virtual server. Advantage of this approach is that almost every application that can utilize mode CPU cores can benefit from this type of scalability without any needs to change the application. Disadvantage is that we have to stop the virtual server to do this configuration change and this type of scalability does not increase an availability of the hosted application. In Windows Azure we are limited because the larges instance has only 8 CPU cores. To scale the application behind this limit we have to use the horizontal scalability. When we scale application horizontally we add more compute nodes and distribute the load between them. This approach works correctly with stateless applications, which are mostly web applications so a load balancer can distribute user requests independently on an inner state of the application on the specific node. Not every application can benefit from the horizontal scalability. Applications that work with the inner state have to be updated to store their inner state in a storage shared between all nodes using Azure Storage, SQL Database or Caching Service. Advantage of the horizontal scalability is that we can add more nodes or remove some of them without an interruption of the service running on existing nodes, simply by reconfiguring the load balancer. We can effectively use the horizontal scalability to adjust the reserved compute power to meet current load requirements. Availability of compute services is at least 99.95% according to SLA with minimum of 2 deployed instances. When one of the instances fails, load balancer detects the missing instance and stops sending requests to the missing virtual server. When using the Virtual Machine service, it depends on the hosted application if supports failover scenarios. 2.3 Storage Services Windows Azure offers specialized services for storing and retrieving data. Windows Azure storage works as a storage for binary data, tabular data (No-SQL approach) and queue messages. Current web services very often use SQL databases, which can be hosted in a second storage service named SQL Database. Windows Azure Storage is a key service for storing data in Windows Azure cloud providing almost unlimited storage capacity. Depending on a type of stored data user can create different storage containers – Blob Storage for storing large amounts of unstructured text or binary data such as images or video. Table Storage offers No-SQL functionality for applications that need to work with large volumes of unstructured data. Data in the table storage are stored as rows in a large table without a defined schema as required in SQL databases. Each row has two important fields that together uniquely identify each row – Partition Key and Row Key. Each row then can contain other fields of basic data types. In such a table there can be more different types of rows. The most important attribute of this storage to consider is that data can be effectively retrieved only based on their Partition and Row Keys. Any other type of search is very inefficient and slow. Queue Storage works as a highly scalable queue that can be used by applications to exchange small messages between different components to create indirect dependencies. Very often this queue is used as a queue for pending task to be processed by worker roles. This approach allows off-load complex calculations from a web role to a worker role, so the web role can serve higher volume of less demanding requests and slow computations are done asynchronously by worker roles, which can be then easily horizontally scaled. Windows Azure SQL Database is a hosted version of Microsoft SQL Server database with minor modifications to provide a high availability by replication data among three synchronous replicas. This database service has advantage in full compatibility of database clients with Microsoft SQL Server. This means that an existing database can be migrated from SQL Server to the cloud with some minor changes and existing applications only has to update its connection strings to point at the hosted version of the database. 2.4 Scalability and availability of storage services Windows Azure Storage guarantees according to SLA availability at least 99.9%. This service has a very complex multi-tier architecture [3] that helps achieve high availability and high scalability of this service. Blobs and Tables are an ISO 27001 certified managed service, which can auto scale to meet massive volume of up to 200 terabytes and throughput for all accounts created after June 7th, 2012. An additional Content Delivery Network service can be used to increase a throughput of Windows Azure Blob Storage. The CDN works as a distributed file cache with servers spread all over the world. When a user requests a file, it is served by the closest CDN node which in case of the first request loads the file from Azure Storage and then holds it for 72 hours in its cache. SQL Database service offers availability of hosted databases of at least 99.9%. Because this service works in a multi-tenant environment where database servers are hosting databases of different customers there are specific policies that limit the maximum number of database connections and the overall load at the database to provide all customers the same quality of the service. To scale this service up there is available a new premium version of this service, which creates a dedicated database servers without mentioned limits especially for the customer. These premium servers are fully manage by Microsoft so there is no additional maintenance effort for the customer. 3 Recommendations to increase service scalability This section describes general principles to increase scalability of a designed service. [4] 3.1 Software components 1. Every component of a software system should have a clearly defined purpose. The web service should be built as a collection of interconnected software components. It is preferable to design smaller components, which are easier to design, code and maintain. Together with increasing size of a component there is also increasing number 2. 3. 4. 5. of dependencies for a given component. It may be very difficult to efficiently identify a performance bottleneck in a very complex component. Every component should have a clearly defined communication interface. It should be possible to simply swap an inefficient implementation of a component for a new more efficient implementation. Having a clearly defined communication interface between software components makes this process simpler and cost effective. Do not burden component with interaction mechanism responsibility. Every software component has its responsibilities and dependencies. A mechanism for communication between components should not be implemented as a part of a component because it increases its complexity and decreases its versatility. Distribute data sources. It is preferable to distribute data sources across different layers of the service. All data of the service can be stored centrally in a single relational database. But together with increasing number of processing nodes there is increasing load at this central database. Relational databases are in general poorly scalable so this central database becomes soon the performance bottleneck we will need to replace. We can use different storage services available in the cloud with different performance characteristics and pricing models to create an efficient and scalable multi-tier storage. Use data replication. We can replicate data to be stored in more than one copy closer to a computation node, so it can access the data faster (e.g. cache a data object in an inmemory cache of a web server). This approach works easily for read-only data, which can be easily without any conflicts stored in many identical copies. Using the data replication for data that are continuously modified by worker roles requires implementing a synchronization mechanism that propagates changes across all nodes in the system. Depending on the target characteristics of this synchronization mechanism, it may be very difficult to implement it. In heavily used systems with many changes in replicated data there will occur many change conflicts that may be difficult to resolve in an automatic way. 3.2 Connection between software components 1. Use explicit connectors. Connectors that are used to connect different components together should not be an integral part of the component so they can be easily replaced by different implementation. This can be achieved by creating a technology-dependent communication interface that wraps over an inner technology independent communication interface of the component. 2. Each connector should have a clearly defined purpose. So far we discussed the distinct purpose of a component but this apply also for communication interfaces which should support only few required protocols and we should rather create more simpler connectors for each protocol than one complex connector. 3. Use the simplest available connector. There are available many different connectors for different communication protocols that support different set of functionality. Developer should choose the simplest possible connector that meets all current requirements instead of a complex connector that may be useful in future. 4. Choose between direct and indirect dependencies. Direct dependencies between components like a synchronous procedure call have benefits in an obvious control flow in the system. Use of indirect dependencies where requests to call procedures are sent over a shared queue to be called asynchronously has better scalability characteristics because they can be distributer among higher number of workers. 5. Do not put application logic into connector. We should clearly distinguish between a purpose of a component and its connector. The purpose of the connector is to mediate communication between the component not to run an additional application logic that would increase connector’s complexity and slow it down. 6. Use scalable connectors. To increase overall scalability of the application it is important to use a scalable communication mechanism between its key components. 4 Architecture of service designed for cloud Based on Windows Azure service’s characteristics and general recommendations for an architecture of the software service I design following high-level service architecture that is depicted in Figure 1. Figure 1: Proposed high-level service architecture 4.1 Storage Tier Based on the recommendation to distribute data sources of the designed service we should use multiple services for storing data so a user load can be effectively distributed among them. Storage services should be tiered based on their scalability and availability characteristics. The service with the best scalability should process majority of user’s requests. The poorly scaling service should process only a few specific requests. This means that we use the relational database service SQL Database to store most of service’s structured data because it provides a mechanisms for transaction processing and can easily ensure consistency of stored data. But we cannot stress the database with all user’s load because of limited scalability caused by the multi-tenant hosting environment. Therefore we use Windows Azure Storage to store cached data loaded from the database in the form that requires minimum processing by compute service to lower its load and allow better scalability of the frontend server. Disadvantage of this approach is a data duplication, but the overall costs for storage capacity in Windows Azure Storage are minimal. We should effectively combine Windows Azure Storage services like the Table Storage where we can store pre-processed data that can be addressed by unique key that is required for effective retrieval of stored data. For binary objects, like images and other resources that we need to distribute to client, we will use the Blob Storage. The client can download files directly from the Azure Storage without need to generate a load on our frontend servers in case files are stored in a public container in the storage. To increase the distribution capacity of this service we use the Content Delivery Network to distribute files that will not be changed often or we can bypass the inability to invalidate a CDN cache when file changes by implementing a mechanism that generates new files with new names after each change. To increase a performance of the storage system we can put the most often used data into the in-memory cache of the web server but we have to deal with the replication problem, because when we cache data that may change, each server can cache a different version and this may lead to inconsistent results when a load balancer lets client communicate each time with a different server in our service. Therefore we can use a Cache Service in Windows Azure that will host our in-memory cache in a farm of dedicated servers. The data retrieval from this cache is a way faster that from Azure Storage so we get performance benefits and at the same time there is not the problem with different versions of cached data. This service is the most expensive type of storage so we should put read-only data to the local cache and use this shared cache only for data that will change. When we use all available storage technologies in the correct combination we can significantly increase scalability of the designed service and also optimize its response time and lower running its costs. 4.2 Compute Tier Clients communicate with the designed service via a HTTP REST communication interface, which is hosted by a farm of web servers (does not matter if the Azure Cloud Service or the Azure Websites Service is used because this API can be hosted in both services and both mentioned offer vertical and horizontal scalability). For resource intensive long running compute operations it is more efficient to use an asynchronous processing. This means that if the operation runs for a considerably longer time than other requests at the web server, it is better not to process the request but to put it into a processing queue. A different server running as a worker can load this operation from the shared queue and process it asynchronously. This approach brings following benefits: 1. Response time optimization. The goal is to lower the overall load at front end servers to decrease the response time for client requests. Therefore we use more storage services as described earlier and we try not to process long-lasting requests that may exhaust compute resources so other requests will have to wait in an internal queue of the webserver for processing. 2. Better load distribution. Because the network load-balancer monitors only availability of the web server and not its current load (due to large amount of monitoring data and requirement for extremely fast request redirection) it can happen that load balancer will redirect a more resource expensive request to a small group of servers that will become overloaded. Other servers in the farm may be idle at the same time. The network load balancer cannot effectively distribute more complex compute operations without a knowledge of the servers load therefore we have to implement this mechanism on our own. At the design time we know what operations will be resource intensive so we can optimize them to run asynchronously on a different server. 3. Recovery after error during processing. It can happen that processing of the longrunning operation fails due to an infrastructure failure or inner exception during processing. Windows Azure Storage Queue and Azure Service Bus Queue both implement two-phase message retrieval from the queue. Worker process loads a message and it becomes for a certain time invisible to others. If is the operation processed successfully the worker tells the queue to remove the message, but if the worker fails the message with a task definition appears automatically back in the queue and can be processed by a different worker. Azure Service Bus Queue also supports moving messages to the specific queue after certain number of failures so they can by for example reviewed by a developer or an administrator. This approach brings also some disadvantages: 1. Asynchronous processing of the task. Basic paradigm of the web and a stateless HTTP protocol is that a client requests data from a server and the server responds with data the client wants. The server cannot contact the client in case of an event at the server side. When we use asynchronous processing in the web environment there is problem that the server cannot send to the client results of the asynchronously processed operation, it can only send a confirmation that this operation was queued. When the operation is finished, there is no way how we can notify the client using a basic HTTP protocol. To resolve this issue we can use a modern approach of a WebSocket protocol. This protocol allows us to have an opened connection with every client so the server can deliver results of the asynchronously processed operation after its completion. It is also possible to off-load traffic to Windows Azure Storage by delivering to the client only an URL address where the results are stored. It depends mostly on the size of results. This behavior has to be implemented on our own and it may significantly increase an overall complexity of the service. 2. Delay in task completion. This approach may lead to longer overall time needed to process the results of less complex operations than in case of the synchronous processing because of a queue service delay, a wait time of the request in the queue and a notification delay. Asynchronous processing of complex compute operations on different servers significantly improves overall service scalability. 4.3 Client Tier The Client Tier is in this paper discussed very briefly because it is strongly dependent on a used technology and each technology has its specific rules for creating well-designed application architecture. The chosen communication protocol for hosted web services is HTTP REST, which is supported by wide range of technologies used for development of client applications. While designing the architecture of the client application it should be kept in a mind, that it should follow the same recommendations as mentioned for the service’s architecture – clearly defined components with defined communication interfaces. When communicating with the server, the client should utilize a client-side caching when possible to decrease the service load. 5 Evaluation of implemented service using load tests I have implemented a reference service according to described recommendations. This service was deployed to the Windows Azure cloud and some characteristics were benchmarked using the Performance Load Testing in Visual Studio. Load tests were running on my server in a datacenter to achieve a higher throughput of the Internet connection. 5.1 Test Case 1: Impact of local cache at service scalability In this test two different implementations of a REST web service are evaluated. The first implementation every time loads data from an Azure SQL Database, the second implementation uses a local cache for data loaded from the database. This service was tested in many configurations of the cloud environment to capture an impact of the vertical and horizontal scalability. Parameters of different virtual server configurations are described in Table 1. Compute Instance CPU (GHz) Memory Storage Bandwidth Extra Small (A0) 1 (shared) 768 MB 20 GB 5 Mbps Small (A1) 1.6 1.7 GB 225 GB 100 Mbps Medium (A2) 2 x 1.6 3.5 GB 490 GB 200 Mbps Large (A3) 4 x 1.6 7 GB 1000 GB 400 Mbps Extra Large (A4) 8 x 1.6 14 GB 2040 GB 800 Mbps A6 4 x 1.6 28 GB 1000 GB 1000 Mbps A7 8 x 1.6 56 GB 240 GB 2000 Mbps Table 3: Available virtual machine configurations in Windows Azure Cost / hour 0,02 $ 0,08 $ 0,16 $ 0,32 $ 0,64 $ 0,90 $ 1,80 $ Load test results for the service without the local cache: Extra Small Small Medium Large Requests per second 2000 1615 1462 1500 1194 969 1000 500 390 144 180 721 706 670 554 1092 487 264 340 294 0 1 2 4 8 Number of web role instances 16 Load test results for the service with the local cache: Extra Small Small Medium Large Requests per second 2500 2105 2000 1676 1500 1758 2021 1841 1653 2074 1805 1648 1554 1725 1264 935 1000 500 2064 558 469 0 1 2 4 8 Number of web role instances 16 Included test results clearly shows the importance of more complex tiered storage, where version of the service with active in-memory cache can process 2-4 times more requests than version without cache. Note: I expect that this service is capable of processing even more requests, particularly in configurations of 8 and 16 instance, and in scenarios with medium and large instances. The performance bottleneck is probably the computer where the tests were running. 5.2 Test Case 2: Scalability of binary file delivery Significant part of the overall service traffic is caused by a transfer of images and other resources. Therefore I wanted to test which way of storing images is more effective and scalable. The first implementation of a tested service loads images stored as binary streams in a database SQL Database. Storing binary files in a relational database is an approach that many current web services use. The second implementation stores data at a local drive of the web role. This approach is suitable only for resources that are part of an installation package of the application. The third approach tests performance of the Windows Azure Blob Storage. The fourth tests performance of the Windows Azure Blob Storage combined with the Content Delivery Network. Results for the service implementation that loads image data from the database: Extra Small Small Medium Requests per second 2000 Large 1648 1624 1500 1407 1261 1710 1292 1210 1029 1000 697 667 500 543 358 352 204 102 49.6 0 1 2 4 8 Number of web role instances 16 Results for the service implementation that loads image data from the local disk of the web role: Extra Small Small Medium Large Requests per second 2500 2000 1901 1731 1706 1870 1749 1679 1810 1718 1959 1916 1801 1500 1084 1000 500 370 529 186 94.1 0 1 2 4 8 Number of web role instances 16 Windows Azure Blob Storage: 1845 requests/second Windows Azure Blob Storage + CDN: 1639 requests/second Test results demonstrate that currently used approach to store binary objects in a relational database is not as effective as use of the Windows Azure Blob Storage. When files are loaded from a local drive of the web role results are much better, but this scenario works only for files distributed together with an installation package. The Content Delivery Network has almost the same performance as the Windows Azure Storage. 6 Conclusion This paper briefly introduces the most important services of the Windows Azure cloud platform and discusses their scalability and availability characteristics so the reader understands that current generation of web services which were built to run on a single web server loading data only from relational database cannot fully benefit from the cloud. To design architecture of highly available and scalable cloud service it is needed to take into account wide range of offered cloud services and discussed recommendations in Sections 3 and 4, which generally describe how to effectively use these services. References [1] MICROSOFT. Windows Azure: Microsoft's Cloud Platform [online]. 2013 [cit. 201311-10]. Dostupné z: http://www.windowsazure.com/en-us/ [2] BONDI, André B. Characteristics of scalability and their impact on performance. In: Second International Workshop on Software and Performance, WOSP2000: proceedings : Ottawa, Canada, September 17-20, 2000 [online]. New York: Association for Computing Machinery, c2000 [cit. 2013-11-10]. ISBN 158113195x. Dostupné z: http://dl.acm.org/citation.cfm?id=350432 [3] CALDER, Brad et al. Windows Azure Storage: A Highly Available. In: SOSP 2011 — Proceedings [online]. 2011 [cit. 2013-05-24]. SOSP '11, October 23-26, 2011, Cascais, Portugal. Dostupné z: http://sigops.org/sosp/sosp11/current/2011-Cascais/printable/11calder.pdf [4] RICHARD N. TAYLOR, Richard N.Nenad Medvidović. Software architecture: foundations, theory, and practice. Hoboken, N.J: Wiley, 2010, s. 468-475. ISBN 9780470167748. Support for Partner Relationship Management Internet Data Sources in Small Firms Jan Ministr VŠB – Technical University of Ostrava Faculty of economics Sokolská třída 33, 701 21 Ostrava jan.ministr@vsb.cz Abstract The paper analyses the processing of data sources to increase competitiveness in small businesses by using of Partner Relationship Management systems. Timely and accurate information about a business partner decisively affect the level of cooperation with a partner and it also reduces the risk of closed shops. The paper describes the possibilities of processing data sources of small businesses in terms of their search and further processing in the combination of information that are stored by in the enterprise information systems. In conclusion, the article discusses the availability of such systems in terms of cost and technology required. Abstrakt Příspěvek analyzuje možnosti zvýšení konkureneceschopnosti malých firem v oblasti Partner Relationship Managementu. Včasná a pravdivá informace o obchodním partnerovi ovlivňuje rozhodujícím způsobem rozsah spolupráce s daným partnerem a snižuje riziko uzavřených obchodů. Příspěvek popisuje možnosti zpracování internetových datových zdrojů z hlediska jejich vyhledávání a dalšího zpracování v kombinaci informacemi, které jsou standardně uchovávány v podnikových informačních systémech. V závěru článku je diskutována dostupnost takových systémů z hlediska jejich nákladů a požadované technologie. Keywords Partner Relationship Management, Data Source, Competitiveness, Resource Description Frame, Web Ontology Language Klíčová slova Partner Relationship Management, datový zdroj, konkurenceschopnost, Resource Description Frame, Web Ontology Language 1 Introduction Partner Relationship Management (PRM) systems support the management of relationships with external, mainly business partners. Alternatively, these systems are also referred to as systems Customer Relationship Management (CRM) or Customer Intelligence (CI). The activities of these systems can be understood as a continuous, ethical process of acquiring and evaluating data, information and knowledge from around the company to achieve a better position in the market [5]. Data sources PRM systems can be divided into the following two categories: internal corporate data sources, which have a stable data structure in form of database records; External data sources include in particular data available in the Internet [1], which have unstructured form and they are primarily stored in: special public portals; corporate portals; social networks, blogs and Internet discussions; social electronic media (e-magazines, e-zines, etc.); e-mail messages, etc. Processing of web data sources can be divided on the basis of the stability of their data structures on processing: Internet data with relatively stable structure that is stored on a special public portals; non-structured texts, which represent additional source of information, however, they can significantly impact in particular the business of the corporation [2]. Processing of external information sources in the context of PRM provides special opportunities for companies whose use of these companies also gets a significant competitive advantage [3]. These mainly include the required information about business partners and customers, competition in order to increase their own sales success and potential employees. 2 Processing of Web Data Sources Data from the Internet Data sources can be obtained in two ways: 2.1 Indirectly (offline, one-off or repeatedly) as a service provided by specialized SW companies. Data are usually structured. However, this method has certain time lag and is very costly. Directly (online, continuously) as a service of a component of the corporate information system RPI. The key problem here is the identification of relevant data sources. In the Czech Republic, there are several public portals providing paid or free information on companies and contacts of their key employees. These data sources include primarily the following: www.ipoint.cz, www.justice.cz, www.statnisprava.cz, www.zivefirmy.cz, www.abc.cz. Processing Web Data Sources with Relatively Stable Data Structure These sources provide current information on the company such as: Name, ID Number (IČO), Tax ID (DIČ), Registered Offices (Street, Postal Code, City), Contact (P.O.BOX, Telephone, Fax, E-mail, Website), The state of the company (for instance, pending execution procedure, bankruptcy, etc.), names and contacts of key employees and officers. The information can be downloaded using open source tools, or other, such as Visual Web Ripper. The tool downloads the required data, converts them to the necessary format and stores them in the corporate database. A drawback of such approach of data processing is the relatively instability of Internet pages and nonstandard description of the displayed details. This fact requires monitoring of the structure of selected data resources and subsequent updating of data stored in RDF (Resource Description Framework), which contains a description of both metadata and procedures for processing them. For more advanced search, Web Ontology Language (WOL) must be used, which extends the features of RDF in the area of creating ontological structures, as is shown, for instance, in Figure 1. The relationship to the company The owner He (she) works for the company The member of the management The employee Figure 1. Ontological Model of Expressing the Relationship to the Company (Source: own) The structure of the described software solution is based on the definition of the Web information sources using object classes (Figure 2), which are specified in more detail for the individual tool used for the downloading of data (for instance, Visual Web Ripper). The class “Source” is specified by three elements defining its purpose [7]: input parameter for searching in the defined resource (“ResourceSearchBy”); type of result for the information obtained from the resource (“ResourceResultType”); identification of the tool used for data downloading (“Tool”). Another class is “Query”, representing every query entered in the system by the user, used to initiate new retrieval of information from the Internet. The class “Result” is the super-class of all result types, it contains the data field Note, used for recording of information that cannot be included in the defined data item in the concrete subclass used for the result type. A subclass of the class “Result” is the class “ResultEntity” defining the data items specific to the entity. Every result of a query execution contains information on the resource that was used and the query in which it was created. A benefit of this processing type is obtaining of up-to-date information from verified resources in the required and clear form of view. Another issue is the need of acquiring “new” knowledge of working with semantic Web, which increases the demands put on the personnel in charge of administering the PRM system. ResourceSearchBy Id. Number Name ResourceResultType Subject of search Visual Web Ripper Query Result Resource Name Priority Tool Note Query time Search text ResultSubject Address Establishment date et cetera Figure 2. Chart of Classes with Relatively Stable Data Structures (Source: own) 2.2 Processing Unstructured Web Data Sources Overwhelming majority of data stored in the Internet has the form of unstructured texts, which must be converted into clear and uniform user environment. The whole process can be divided into the following 4 stages [6]: monitoring and collection of source data; administration and storage of source data; data processing using analytic tools; presentation of analytic results. During the process of processing unstructured data for the aforementioned purposes, one must first define the target of the Internet monitoring, in our case, this is an individual or a company we wish to monitor. Further, access of the defined social networks and discussion forums must be obtained. Here the current laws of the relevant country must be observed, for instance, not to pose as someone else, etc. Selected targets are then monitored in the Internet and if the search engine finds any “interesting” information on the target, it is stored on the local server in a smart database, which allows, first of all, fast searching and data indexing. This data is subsequently analysed and the results of the analysis are reported to the end user. On the level of selected text database, analytic functions sorting data by time, source, type, etc., are normally used. On this level of unstructured text processing, analytic functions such as filtering of selected word or a person's name can also be used. These operations are usually carried out during later processing stages using special software tools, which must be calibrated according to domain areas, which is not usual for the database-level tools [8]. Processing of unstructured data can be divided into the following areas: content analysis; entity identification (authorship analyses); relations analysis; sentiment analysis. Content analysis is one of the main areas of unstructured text processing. The goal is to identify key words that appear in the text. However, to achieve this, words carrying semantic information, i.e., combinations of selected nouns and adjectives, verbs and numerals, are selected from the text. At the same time, the synonyms that make terms equal in meaning replaced with one selected proxy. Such modified text is then used for deriving the topics of the different texts, which can have, for instance, the form of a list of keywords. The whole process of content analysis consists of several main steps that are described in the following paragraphs [4], [10]: Tagging of word types using external libraries (such as POS Tagger). From such tagged words, keywords are selected, mostly nouns, verbs and adjectives. Processing of synonyms permits grouping of words according to their meaning, not only depending on their form, but also with an option of creating relationships between them. Lemmatization and stemming tries to convert words to their basic forms. Lemmatization is more time-consuming as the processing is based on the context of the word. Words found in different forms must be recognized as identical. Then equivalent words can be grouped and it is possible to avoid multiplicity of keywords [19]. Person Employe r Interests Educatio n Squash Bicycle Motorcyc le Motorcycles Motorbike Motorbikes Bike Other synonyms Figure 3. Example of Ontology (Source: own) Ontology of decision-making trees is a different and simplified approach to processing of analysis for unstructured texts from the one described above. It is based on the pre-definition of keywords, i.e., entity attributes, in advance by the user. From these attributes, a query in the decision-making tree is generated, with the nodes of the tree being the queried entity attributes (areas of information needs of the user with respect to the entity), which are expressed by a table of synonyms for the queried area. The detail level of the query is then expressed by the level of explosion of the given node, see Figure 3. The structure of the query can be optimized using the decision-making trees induction theory or Occam's razor [2]. A shortcoming of this approach is the fact that in generation of the list of entity attributes, not all key attributes need be specified that can be identified using analytic tools directly from the unstructured texts than the approach described above, for instance, we monitor attributes such as sports for some persons, but we cannot find out from the given data that he has started gambling. In general, this approach is appropriate in the segment of small business as the application development is less costly. 2.3 Presentation of Analysis Results The results of the analyses are presented in three different forms: detail of the found record; graph, which is used primarily for visualization of the development in time of the phenomenon (line and bar graphs), visualization of percentages for the different entity types (pie-charts), or for visualization of relationships (network charts, polygons); combination of graphs, i.e., special type of visualization such as the spectrum graph, which is often used in combination with Liferay, a corporate portal based on Java EE, allowing easy integration of information, applications and processes. The portal presents information to its visitors via its Web interface, which can be customized to suit the current needs of different users. The form of presentation of results is a very important part of the application, and it basically sells it, because the user must quickly comprehend the information in a wider context. Figure 4. Example of Data analyses Presentation (Source: own) 3 Specifics of the PRM Add on in Small Business In implementing the PRM add-on in small business companies, one must take into account the purpose of the future utilization of the module. Most current users in this segment will not practically use all the options mentioned above, and at the same time, the experience of the author of this article suggests that they do not know which type of information they precisely require and in what form they would like to obtain it. Therefore, one must very carefully regulate sometimes unrealistic ideas of the users that all they would like to know is really freely available in the Internet, and at the same time, advise them on the costs of obtaining such information. Most of the current solutions that are offered on the market solves this problem off line approach or the processing of relatively expensive one-off analyzes, that not cheap. For these reasons, the author recommend very precise identification of the domain areas and the type of information resources in the Internet. On completing a serious presentation of the options and costs of the PRI module add-on, most customers will be satisfied with their information system featuring online processing of Internet resources with a relatively stable data structure, where within the small business segment, most users require basic information on the business partner or customer (registered offices, economic condition). Processing of unstructured data from the Internet allows for obtaining more “specific” information on the entity, and one can be better acquainted with the characteristics of an “unknown” business partner or with new information on existing entities, however, it must be considered that every such solution is fundamentally unique and therefore, at present, it is more costly. In the small business segment, most users require standard information on their business partner or customer. It is very rarely that these customers want processing in the sentiment analysis area, as they are mostly in direct contact with their customers and know their opinions. Further, one must take into account the technological limits of the companies and therefor, development of such applications will probably be directed towards using a cloud computing platform. 4 Conclusion Currently, it is possible to support the small business in the area of obtaining customer data, using the approach combining data from the primary ERP system with a PRM add-on. For processing of data using the PRM module, practical options are structure data (or data with a relatively stable structure) and unstructured data. For deployment in small companies, it is important to define the purpose of processing such data. However, what is decisive for the deployment of PRM, is the cost of prepared application, as at present, no typical solutions are on the market. As for the requirements of standard information, offline analysis can be used, and it is possible to make a selection from the companies processing data on economic entities on the domestic market. In the case of more complex requirement with respect to the volume and structure of processed information, the turn-key solutions would entail large financial burden. 5 Acknowledgements This project is based on the experience obtained in solving the OPPI project of the Czech Ministry of Industry “Use of ontology and semantic Web technologies in a CRM solution”, in which the approaches of Internet information resource processing described in the article were applied. References [1] D. M. Boyd, N.B. Ellison, N. B. Social Network Sites: Definition, History, and Scholarship. Journal of Computer-Mediated Communication, vol. 13, no. 1, pp. 210-230, 2007, doi: 10.1111/j.1083-6101.2007.00393.x. [2] F. Nguyen, T. Pitner, T. Information System Monitoring and Notifications Using Complex Event Processing. Paper presented at the 5th Balkan Conference in Informatics BCI. Novi Sad, Serbia. pp. 211-216, 2012, doi:10.1145/2371316.2371358. [3] J. Ministr, J. Ráček. Analysis of Sentiment in Unstructured Text, In IDIMT-2011: Interdisciplinarity in Complex Systems. pp. 299-304, Jindřichův Hradec: Trauner Verlag, Linz 2011. [4] V. A. Yatsko, T. N., Vishnyakov. A method for evaluating modern systems of automatic text summarization. Automatic Documentation and Mathematical Linguistics, vol. 41, no. 3, pp. 93–103. 2007, doi: 10.3103/S0005105507030041. [5] J. Hančlová, P. Doucek. The Impact of ICT Capital on Labor Productivity Development in the Sectors of Czech Economy. In IDIMT-2012: ICT Support for Complex Systems. pp. 123-134, Jindřichův Hradec: Trauner Verlag, Linz 2012. [6] ALLEMANG, Dean a James HENDLER, 2011. Semantic Web for the Working Ontologist: Effective Modeling in RDFS and OWL. 2nd ed. Waltham: Morgan Kaufmann Publishers, ISBN 978-0-12-385965-5. [7] HEBELER, John, Colin EVANS a Jamie TAYLOR, 2009. Semantic web programming: modeling in RDF, RDFS, and OWL. 1st ed. Indianapolis: Wiley Publishing, 616 s. ISBN 978-047-0418-017. [8] YU, Liyang, 2007. Introduction to the Semantic Web and Semantic Web Services. 1st ed. Chapman and Hall/CRC, 368 s. ISBN 978-1584889335. [9] Christopher D. Manning Yoram Singer Kristina Toutanova, Dan Klein. Feature-rich part-of-speech tagging with a cyclic dependency network. In Proceedings of HLTNAACL, pages 252–259, 2003. [10] Stefan Evert Eugenie Giesbrecht. Is part-of-speech tagging a solved task? an evaluation of pos taggers for the german web as corpus. In Web as Corpus Workshop WAC5, pages 27–35, September 10th Summer School of Applied Informatics Proceedings 10. letní škola aplikované informatiky Sborník příspěvků Bedřichov, 6.–8. září 2013 Editoři/Editors: prof. RNDr. Jiří Hřebíček, CSc. ing. Jan Ministr, Ph.D. doc. RNDr. Tomáš Pitner, Ph.D. Vydal: Nakladatelství Littera 2013 1. vydání Tisk: Tiskárna Knopp, Černčice 24, 549 01 Nové Město nad Metují ISBN 978-80-210-6058-6