IS-270, Week 2
The database development
lifecycle and fact-finding
Jean-François Blanchette
Department of Information Studies
Graduate School of Education and
Information Studies
Components of a Database System
Organizational Database Systems
1
Database Management System
ο A database is a self-describing collection of related
records
ο A database management system (DBMS) serves as
an intermediary between database applications and
the database
ν
It creates, processes and administers the databases it
controls
ο A database application is a set of one or more
computer programs that serves as an intermediary
between the user and the DBMS
ο Users use a database application to enter, read,
delete and query data and to produce reports
Functions of a DBMS
ο
ο
ο
ο
ο
ο
ο
ο
ο
ο
Create databases
Create tables
Create supporting structures
Read database data
Modify database data (insert, update, delete)
Maintain database structures
Enforce rules
Control concurrency
Provide security
Perform backup and recovery
Real-world DB design is difficult!
ο The implementation of a information
processing/management system is a
significant engineering endeavor:
ν Project must complete on time/on budget;
ν When operational, system must meet its
requirements, operate reliably and efficiently;
ν Documentation and coding for the project must be
such that the system can be maintained and
enhanced over a long lifetime.
ν Most importantly, it must meet the needs of its
user
2
Very difficult!
ο “The software crisis”: 1996 UK study states
that:
80-90% systems unable to meet their goals!
80% delivered late/over budget
40% projects fail/abandoned
Less 25% properly integrate business with
technology objectives
ν Only 10-20% meet their success criteria
ν
ν
ν
ν
ο Two major causes:
ν (A) Poor or incomplete requirements specification
ν (B) Lack of structured development methodology
(A) Poor requirements
ο One explanation: systems analysts trained to
focus on considerations of technical efficiency
(speed, storage, cost)…
ο … not on organizational aspects of information
systems implementation,
ν How the system will fit within a given
organizational structure
ν How the system will fit the mental structures of
users (cognitive ergonomics)
ο Fact-finding methodologies have evolved in
response to this problem
(B) Steps to structured development
ο
ο
ο
ο
ο
ο
ο
ο
ο
ο
(1) Planning
(2) System definition
(3) Requirements collection and analysis
(4) Logical design
(5) Physical design
(6) Application Design
(7) Prototyping
(8) Implementation
(9) Testing
(10) Maintenance
3
Database planning
ο Mission statement and objectives:
ν Defines the major aims of the DBMS
ν Identifies particular tasks DBMS must support
ν Estimates resources: work, money, time
ο Documents how data will be collected
ο Documents legal and business requirements
regarding data acquisition (feasibility?),
manipulation (privacy?), storage (security?)
ο Other contexts to consider: organizational
culture, ethical practices, political issues, etc,
may all play a role in success of project
System definition
ο System boundaries: How does the database interface
with other systems within the organization?
ο User view: What is required of a database from
perspective of particular users/departments. F.ex., for
a library catalog database:
ν
ν
ν
ν
Patrons needs to perform searches/browse through the
catalog to view description of library material;
Clerks needs to enter new patron information; check out
books, print out receipts for late fees;
Branch manager needs to enter new employee information;
Regional manager needs to extract figures for all branches
within the province
Requirements
ο Collect information for each major user view
using fact-finding techniques:
ν Data to be used and/or generated
ν Expected transactions (retrievals, updates)
ν Policies for each views (what can and can’t be
done with the data)
ο Analyze and distill information into statements
regarding expected functionality of systems
ν Patron: perform searchers by title and author; …
ν Clerk: enter patrons, etc…
4
Database design
ο Logical design:
ν identify and model the objects to be represented
ν Identify and model their attributes, relationships,
constraints
ο Physical design:
ν Translate model into relational tables
ν Decide file organization
ν Decide indexes
Application design
ο In our case (Access), application means queries
(transactions) and the input/output screens (forms
and reports);
ο Three types of transactions:
ν
ν
ν
Retrieval of data for display or report
Update: insert, delete, or modify records
Mixed: retrieval and update
ο For each transaction, document:
ν
ν
ν
Data to be used by transaction
Output of transaction
Importance and expected rate of usage
ο Good and friendly user interface design is crucial to
usability
Prototyping, etc.
ο Prototyping: create working mockup, in order
to test and refine design
ο Implement: physical realization
ο Testing:
ν Debug
ν Verify conformance to requirements
ο Operational maintenance:
ν Performance monitoring and tuning
ν Upgrading/update, as new requirements surface
5
Discussion: TR Commission
ο Is it possible to fit emotional & violent events
into cold analytical structure of databases?
What kind of stories do databases tell?
ο Was the data model effective in capturing
'real-world' relationships between victims,
events, violations?
ο What is the relationship between (a) keeping
track of individual violations, and (b) revealing
trends?
ο Advantages/Disadvantages of recording
violations using free-form narrative text, forms,
and semi-structured forms?
Fact-finding
Why fact-finding?
ο In order to properly identify requirements…
ο … and thus derive an accurate data model…
ο .. the designer must learn about problems,
opportunities, constraints, terminology,
priorities of the organization and users
ο Crucial issue! if we correctly identify and
model information management needs, we
will:
ν Make better information systems
ν Guarantee user/market acceptance
ν Even create new markets!
6
Fact-finding techniques
ο Information for the stages of database
planning, system definition and requirements
collection and analysis is obtained through
fact-finding techniques
ο Common methods:
ν
ν
ν
ν
ν
ν
(1) Review documentation
(2) Questionnaires
(3) Interviewing
(4) Participant-observation
(5) ‘Naturalistic enquiry’
(6) Good old-fashioned research
(1) Reviewing documentation
ο Overall information management problem is
found in:
ν Memos, e-mails, minutes of meetings;
ο Description of organization:
ν Organizational charts, business/strategic plans,
job descriptions, forms, reports (manual and
computerized)
ο Current information system documentation:
ν flowcharts, diagrams, program documentation,
user/training manuals
(2) Questionnaires
ο Advantages:
ν Gather facts from large number of people while
controlling the format of answers
ν Confidential, anonymous, convenient
ν (fixed-format) easily tabulated and analyzed
ο But …
ν Preparation is time-consuming
ν Body language of respondent unavailable
ν No opportunity to reword questions when they are
misinterpreted
ν Low (5-10 %) response rate
7
(3) Interviews
ο Most commonly used method: meet face-to-face with
relevant staff and management, using
(semi-)structured format
ο Advantages:
ν
ν
ν
Fine-grained, flexible, allows for rewording
Body language of respondent
Involves end-users in design process
ο But …
ν
ν
ν
Very time-consuming
Costly to transcribe, code, and process …
Requires good communication skills from system analyst
(4) Participant-observation
ο Participate in/observe staff perform job tasks
ο Advantages:
ν Will enable you to observe tacit knowledge, not
explicit by other means
ν Makes visible interpersonal and organizational
dynamics
ο But:
ν To obtain representative observations, must be
performed over significant time period
ν Heisenberg Principle: observation alters
measurement
(5) ‘Naturalistic inquiry’
ο Productivity paradox and emergence of collective/cooperative context for IS leads to growing appreciation
for ‘naturalistic enquiry’
ο Suite of research methodologies/data collection
techniques:
ν
ν
Participant observation, respondent interviewing, artifact
collection, generation of field notes
From anthropological tradition of ethnography & sociological
tradition of ethno-methodology
ο Better at discovering (often tacit) meanings and
understandings that participants in social contexts
negotiate and derive from interactions with IT and
other participants, increasing likelihood that system
will be used effectively
8
Example:
ο In “The myth of the paperless office” Selen & Harper
(2002) use ‘naturalistic enquiry’ to examine how office
personnel use paper to organize their work:
ν
Fine-grained analysis of reading practices (reports, memos,
etc); observation of filing practices; etc.
ο Findings:
ν
ν
Paper intimately tied to how knowledge workers organize
their work, individually and collectively
Replacing all paper with IT is inefficient: better to rethink
work organization and design information systems (with or
without paper) around it
(6) Old-fashioned research
ο Other people might have solved the same
problem you are facing
ο Search:
ν reference books
ν computer trade journals
ν Web
ο F.ex., generic data models are available for
many common information management
scenarios:
ν lending, invoicing, scheduling, grading, etc.
Outputs of fact-finding
ο Collected data is used to produce several
descriptions of future database:
ν
ν
ν
ν
ν
(1) Mission statement
(2) Objectives
(3) User views
(4) Requirements specification
(5) System specification
9
(1) Mission Statement
ο Concise summary of the overall purpose of
the database (e.g. Making the Case):
ν “Telling the truth in such a way that it cannot be
denied is the first need of a truth commission
established in the aftermath of gross human
violations.The purpose of the TRC information
management system is to provide a collective
memory for those violations and the ability to
relate information from different sources. By so
doing, it allows anyone in the organization to
access information collected by any investigator,
without restriction.”
(2) Objectives
ο High-level description of the expected
functionalities of the DB:
ν (a) Maintain (enter, update, delete) data on
branches, staff, material, patrons, borrowing,
suppliers, orders to suppliers
ν (b) Perform searches on material, staff, borrowed
material, members, suppliers
ν (c) track status of book orders, stock, borrowed
material
ν (d) Report on staff, borrowed material, patrons (at
each branch), suppliers, orders (system-wide)
(3) User views
ο Identify classes of users according to
objectives:
ν Director: To report on video, staff, members at all
branches
ν Manager: To maintain (enter, update, delete) data
on branch and staff at given branch
ν Clerk: To maintain (enter, update, delete) data on
videos and members at given branch
10
(4a) Requirements specification
ο For each user view, establish:
ν (a) What data is needed + constraints
ν (b) Transaction (entry, update, queries)
requirements
ο (a) Data requirements (Branch view):
ν Before borrowing, patrons must register as a
member of a local branch
ν Must provide first and last name, address
ν Each member assigned a unique number across
all branches
ν Members can rent up to 10 books at any one time
(4b) Requirements specification
ο (b) Transaction requirements:
ν Data entry:
ο Create new member
ο Create new borrowing agreement
ν Data update/deletion:
ο Update/delete member of staff
ο Update/delete details on book
ν Data queries:
ο List name, position, salary of staff at a given branch,
ordered by staff name
ο List the details of all copies of a given book at a
specified branch
(5) System specification
ο How much data will the database contain?
ν
20 000 titles; 400 000 copies; 10 branches; 2000 staff; 100
000 patrons
ο How much will the database grow by?
ν
100 new titles, 20 staff, 1000 patrons/month
ο What types & average number of record searches:
ν
ν
Catalog searches: 10 000/day (peak 11 am)
Borrowing: 5 000 (5 pm)
ο Performance (response time): for which transactions
is good performance critical to the smooth operation
of the organization?
ο Networking issues, backup, recovery
ο Security:
ν
determine privileges appropriate to user view
11
Assignments
ο Labs:
ν
To make sure you absorb the relational DB concepts
presented in class (relationships, queries, modeling)
ο Requirement analysis:
ν
ν
To make sure you understand the difficulty of analyzing and
solving a data management problem using relational
databases
To increase your familiarity with this genre of technical
writing
ο Database project:
ν
Get to design a reasonably complex DB
ο Final paper:
ν
Knowledge + writing & analytical skills
12