Intelligent Information Systems 4. Electronic Commerce Gio Wiederhold EPFL, April-June 2000, at 14:15 - 15:15, room INJ 211 7/26/2016 EPFL4E - Gio spring 2000 1 Schedule for Seminar Course on Presentations in English -- but I'll try to manage discussions in French and/or German. • I plan to cover the material in an integrating fashion, drawing from concepts in databases, artificial intelligence, software engineering, and business principles. 1. 13/4 Historical background, enabling technology:ARPA, Internet, DB, OO, AI., IR, XML. 2. 27/4 Search engines and methods (recall, precision, overload, semantic problems). 3. 4/5 Digital libraries, information resources. Value of services, copyright. 4. 11/5 E-commerce. Client-servers. Portals. Payment mechanisms, dynamic pricing. 5. 19/5 Mediated systems. Functions, interfaces, and standards. Intelligence in processing. Role of humans and automation, maintenance. 6. 26/5 Software composition. Distribution of functions. Parallelism. [ww D.Beringer] 7. 31/5 Application to Bioinformatics. 8. 15/6 Educational challenges. Expected changes in teaching and learning. 9. 22/6 Privacy protection and security. Security mediation. 10.29/6 Summary and projection for the future. • Feedback and comments are appreciated. 7/26/2016 EPFL4E - Gio spring 2000 2 Definition Sales of Information and goods over the Internet • Information -- two-way transport via the Internet Dig – some backup by FAX, post, express Lib • scientific, health information, training, catalogs, music, videos H A R D E R • Goods -- ordering, often payment, over the Internet – delivery by pickup or shipment – Vouchers -- no physical shipment of goods • tickets for travel, entertainment -- pickup based on voucher – Fungible -- shipped: items are indistinguishable, returnable • books, CDs, oil, building materials an increasing fraction – Unique (to some degree): best viewed before acceptance • cars, groceries, jewelry, art + other dimensions: price of good, time value, spoilable, payment. 7/26/2016 EPFL4E - Gio spring 2000 3 Size and Growth • • • • 1999: 2003: 2006: 2013: $138B *. === small percentage (1%?) of world business $540B * (~ 45% / year) --- $4 000B** ~factor 10 over 1999 if it continues at the current rate* ~factor 100 of current = all unlikely, but likely largest fraction Difference by area – Easy areas sooner -- travel, books, CD, toys • traditional businesses feel the effect now • eventually structural changes » example: programming structural changes * per Yankee Group ** per Business Week 7/26/2016 EPFL4E - Gio spring 2000 t? 4 Business practices Compare to B&M (Brick-and-Mortar) stores • Getting to the customer – portals • general, clientele -specific – follow-up differs a bit by type of good • Getting the customer’s commitment – pre-payment, esccrow, post-payment? • Building – build-to-order is preferable – personalize, low inventory -- less waste, but delay? • Shipping – by own (limited locale) or by common carrier • Return policy – return needed? by carrier (awkward) to B&M if possible more on these topics 7/26/2016 EPFL4E - Gio spring 2000 5 European/US differences • Credit cards -- help E-commerce in the U.S. – more prevalent in US – no / low liability in US – allow canceling payment if goods are unsatisfactory • Store hours -- encourage E-commerce in Europe – long opening hours are common in US, Asia, not in Europe – working individuals, couples driven to purchase on the net? • Sales Taxes -- help E-commerce in the U.S. – Interstate Sales taxes hard to collect (sales taxes may disappear ?) – VAT is invisible, hard to avoid • Store density? Traffic? Security? 7/26/2016 EPFL4E - Gio spring 2000 6 Implementation Many suppliers now for various types of servers • Low density boutique systems ( < $5 000) • fixed rat Internet costs to some limit • Server services (incremental costing, high reliability) • Shared staffing • Portal-based systems • Amazon associates • Dedicated high end systems (10-200+ processors) • in-house staffing • Distributed (peripheral) service replication [Akemai] • reduced bandwidth cost, increased response 7/26/2016 EPFL4E - Gio spring 2000 7 Access to Sales Info. • Currently mainly HTML/CGI scripts to databases – well established technology with many tools, people • Assumption moving to XML-based access unclear – better (content semantics sensitive) format – DTD standards for various domains – full exposure of contents - may not be wanted by suppliers • loss of competitive information, say inventory • selective access for known/registered customers in B2B – XML has representation defined, not yet processing • much initial focus on document processing HTML • no algebra, no universal quantification, no typing – the basis for and computation & composition 7/26/2016 EPFL4E - Gio spring 2000 8 XML access technology • in hold, waiting for W3C blessings of query language: – search, views, joins, sorting, ranked results, keyword search HTML Alg UQ Srch Vw J Srt Rnk Key support XSL A - - C - - B - - [Microsoft] XQL - D A C - - B - - W3C document XML-QL - C - B C B B - - [AT&T INRIA] - B D B [Poli Milano] C D B XML-GL WML* C - C Lorel A B - B C B B [Stanford projects] ... WebML[torisoft] can translate WML to HTML * Many research questions - logical and performance 7/26/2016 EPFL4E - Gio spring 2000 9 Other Functions • Visual GUI Interfaces [XML-GL] • DTD / Schema management (XMI [OMG], RDF, DAML) • required for multi-domain integration • Update capability [Lorel] maintainer vs. querier • Validation capabilty LINT maintainer • Protection maintainer vs. querier types • related to views? Can / should a single language handle all? • SQL doesn’t handle everything • if not, must have a convenient API for common languages C++ • multi-level, mediated architectures (next Friday at EPFL) • insert application with specific domain functions 7/26/2016 EPFL4E - Gio spring 2000 10 Customer Aquisition • Through Portals – pay for each click through – pay for each purchase – pay for each registration show 8% growth rate versus 1%* • Direct mail • Advertising Follow-up with adequate services 42%* poor Corporate presence - public relations 36% inadequate Top level catalog of information 22% satisfactory On-line ordering + with return policy 15% good On-line payment + security 3% excellent Personalization + rapid response * abstracted from1999 IBM survey of 2373 companies 7/26/2016 EPFL4E - Gio spring 2000 11 Personalization • of Interfaces • per individual • per group and assignment of individual to groups – automatic (hard) – by self selection (awkward, unreliable) • per multiple groups – appreciated – privacy concerns - small, vocal group? » “3/4 employees don’t mind email monitoring [AngusReid]” • Vendors: IBM, ATG [Dynamo], Netscape, Microsoft • of Domain vocabularies – DTDs for • universal use [CommerceOne] • B2B legal, chemistry, genomics, oil, . [industry consortia] 7/26/2016 EPFL4E - Gio spring 2000 12 Document Type Definition (DTD) Computation requires is aided by having types – XML is still focused on documents • DTD allows definition of structure, but not of type • always text: PCODE (= UNICODE 8 - 16 bit) • numeric values require additional conventions ( , . E $ £ ) • Additional proposals – XML-schema[W3.org]: to cover all {OO,R}DBMS-like types • extensible base types, typed links, integrity constraints – ambitious – XMI [OMG]: for interchange – WebML[Poly Milano]: web site specification with functions 7/26/2016 EPFL4E - Gio spring 2000 13 Health-related sites • Long-term leverage -- most health states persist • Corporate benefits per person insurance ~$5000/year – early intervention – coercive? • Real privacy concerns – info sold to • pharmaceutical companies • drug retailers • Health Maintenance Organizations – “Employers have information now to weed out high cost employees” 7/26/2016 EPFL4E - Gio spring 2000 14 Communities • Real communities – Example: MaplewoodOnLine.com • Virtual communities – distributed hobby, . . . . groups – 1000’s – distributed problem groups - social, health, . . . – 1000’s 7/26/2016 EPFL4E - Gio spring 2000 15 Portal Payment Schemes • User pays • Free View - public service • Free View - advertisement – Goodwill – pay-per-page or ad display hit (newspaper model) • do robots/ multi-fetch page count? • hacker schemes? – pay per click-through on ads – pay per lead (name, email, category, location) – pay per sale (%tage): Amazon associates 7/26/2016 EPFL4E - Gio spring 2000 16 Google model • • • • • No messy ads on first page when a search is done, then relevant ads may appear + payment by other portals, as Netscape Netcenter + sales to Intranets, corporate sits (as Redhat.com) Growth – – – – – – – – – – – – 7/26/2016 spring 1995 - Page & Brin met as CS PhD admittees Fall 1995 work on Stanford’s DL project fall1996 - Google project disclosure to Stanford OTL summer 1997 - stanford.google.edu services 1998 Page Brin buy a terabyte of disk ($15 000)on their credit cards summer 1998 Angel fundng (Bechtolsheim, Cheriton), October 1998 rent garage from friend, hire Jan 1999 move to commercial space in Palo Alto June 1999 Venture funding $25M. September 1999 Offices in Mountain view September 1999 3.5M searches/day December 1999 6.M searches/day EPFL4E - Gio spring 2000 17 Paying for SW Services You can not run an effective (SW) business and not be reimbursed for it. How? Four approaches: – Sell Software – Lease copy / usage rights – Time / user limited access – Charge by use instance General problems, effects differ – IP protection? protect – keeping SW updated update – billing for est.value bill – performance effect perform June 1998 CHAIMS sell oilfield to customer lease well fill tank provide bus Buy Lease Limit Use poor some fair good poor ok good good simple simple awkw. hard no no little some 18 JMC: support electronic journals by page charges. 7/26/2016 EPFL4E - Gio spring 2000 19 E-Commerce in the Digital Library Steven Ketchpel & DL Economics Group Payment Delivery CyberCash DigiCash First Virtual SET Cryptolope DigiBox HTTP E-mail Major Integration Problem Shopping Models: Pay-per-view, Subscription, Session, Shareware, Auctions, Site License, Gift Certificate, Layaway, Pre-paid vouchers, … . 7/26/2016 EPFL4E - Gio spring 2000 20 Methods • • • • • Trust Escrow Subscriptions Debit cards Credit cards – cost per transaction about $0.50 > many Internet Txs – limited number of acceptors • new scheme: • Wallets Limited amount instantly available, not secured against loss, suitable for small payments, could be anonymous. Poor acceptance. • dynamic pricing 7/26/2016 EPFL4E - Gio spring 2000 21 DL Copyright Management unique id Library of Congress Copyright Office Rights System Work Repositories Authors Publishers Rightsholders courtesy of CNRI requests exclusive rights use rights Readers Libraries Consumers Gio Wiederhold 7/26/2016 EPFL4E - Gio spring 2000 22 Shopping model: merchant-independent logic controlling flow of business model Example shopping models: Order, Pay, (Deliver 52 times) (1 month; Order, Deliver) Pay State Information Event Handlers Abstract API allows application to interact with many different services in a consistent way 7/26/2016 2 1 Order Complete 3 Start Transfer $ 4 Payment Complete Event Handlers Event Handlers Customer Bill Merchant Event Handlers Payment/Delivery/ Other Services EPFL4E - Gio spring 2000 Proxy event handlers translate from native applications to shopping model defined protocols 23 B2B Commerce Total $ B2B volume is 3-5 x retail -- but fewer transactions • XML common business Library [Commerce One] – – – – catalog interoperability to come: bank clearinghouse transaction fee for buyers and suppliers value is in number of potential suppliers / buyers • Specialized consortia [ENERVA for petroleum prods] – focus on automating back-office work for existing types of product exchanges – started by Ethyl corp, 60 participants • Managed services [Envive] Supplier entry requires trust, not lowest price 7/26/2016 EPFL4E - Gio spring 2000 24 Source of Changes If you're listening to your customer. it's almost preordained that you'll miss the new market. And when the new market expands to encompass the old market ... that's when companies can become obsolete [David Isenberg, AT&T researcher, 1998] • Sequence 1. 2. 3. 4. 5. 6. 7. novelty investigation (by many, elation by inventor) disappointment at problems, speed of acceptance market visibility broad acceptance commodization stasis market size no. suppliers 7/26/2016 EPFL4E - Gio spring 2000 25 Conclusions • Much change expected – uncertainty – hard to predict time scales • Industry & Business ahead of academia – Industry has the problems, ad hoc, fast fixes – Most academics don’t listen enough to the world • Standards provide needed infrastructure for next phase – standards require acceptance, not fanciness • Governmental / Organizational initiatives lag Ind. & Aca. – HTML versus 7/26/2016 W3C initiatives EPFL4E - Gio spring 2000 26