Theory of Databases

1
Theory of Databases

• Module lecturer Dr Meurig Beynon, Room 3.15
• Module Content
• Generalities on databases
• Database models Theory and Practice
• Entity-relationship models
• Relational models algebraic and logical
• Commercial relational languages

2
CS319 Theory of Databases

• Module Content (cont.)
• Technical Issues and Algorithms
• Optimisation
• Integrity constraints
• Database design
• Future Trends Theory and Practice
• Object-oriented models
• Extended relational systems
• The Relational Model and SQL

3
Theory in Computer Science?

• Theory in computer science is typically viewed as
•  
• - associated with formal foundations for
  computing
• - based on mathematics (can there be any other
  kind of theory?)
• - used to justify and to inform practical
  principles

4
Common perceptions about theory of CS

• Computer Science means accounting for computing
  practice using mathematical theory
• Mathematical theory is the essential foundation
  practical computing is its application
• The mathematical foundations of computer science
  are the bedrock of the subject - they frame
  computing in an abstract manner independent of
  technologies, and how these abstractions are
  realised in practice is a relatively ephemeral
  concern

5

• Ephemeral conveys a number of ideas
• temporary
• reflecting a current situation or fashion
• not of abiding interest
• of subordinate significance

6
In support of this view of theory

• From a mathematical perspective
• have a powerful abstract theory of computation
  all programming platforms are Turing equivalent
  in some fundamental sense
• it is difficult to know what meaning to attach
  to computational activity for which there is no
  abstract mathematical model
• computation has - and always has had - a crucial
  role in mathematics itself

7
In support of this view of theory

• From a broader scientific perspective
• It is possible to regard pure sciences such as
  physics as establishing an intimate association
  between mathematical models and our interaction
  with the natural world
• CS might plausibly be viewed as an embryonic
  Science of the Artificial, and we might expect
  abstract mathematical models to be ideally suited
  to framing such a discipline

8
In support of this view of theory

• From a cognitive perspective
• reasoning is an important constituent of human
  thought - logic is seen by many computer /
  cognitive scientists as the essential constituent
  of human intelligence, and the computer as an
  appropriate metaphor for the brain
• cf. the logicist perspective on artificial
  intelligence

9
Contrary to this view

• Theory has yet to deliver programming tools and
  principles that meet practical computing needs
  comprehensively
• Functional programming
• Formal models of concurrency
• Formal methods for software development
• represent a partial solution to practical
  issues

10
Contrary to this view

• Crucial aspects of practical computing have no
  satisfactory supporting formal theory
• human-computer interaction
• end-user programming
• object-oriented modelling
• object-oriented software development
• mathematical models are of limited value

11
Contrary to this view

• Some of the most significant aspects of computing
  practice are pre-theory and subjective in
  character
• Consider the observation and experiment involved
  e.g.
• in requirements capture
• in customising interfaces for personal use
• in software development based on prototyping
• implicit knowledge embodied in interaction not
  explicit in mathematical models

12
Contrary to this view

• Interaction in the world can only be viewed as
  accounted for by theory if we accept a
  reductionist stance on the nature of science
  consider the implications / meta-implications of
  a Theory of Everything in physics
• Many aspects of human activity and interaction
  in the world arguably involve more than
  applications of logic interaction in the world
  can be viewed as conceptually prior to reasoning,
  and not explicable in terms of logic consider
  e.g. Rodney Brooks's robotics research on themes
  such as Intelligence without Reason, Intelligence
  without Representation

13
Contrary to this view

• Modern developments of computing challenge the
  mind-body dualism that underlies the idea of the
  brain as a complex computational device that can
  be conceptually separated from its embodiment
  consider e.g.
• - brain-mediated communication
• - meaning as situated
• - context dependence in natural language
• Mathematics does not account for situated meaning

14
Two possible approaches to developing CS

• evolutionary accept that there is an essential
  duality within computing that distinguishes the
  hard core science of computing from the soft
  ephemeral pseudo-science of computing practice,
  and focus on developing and extending classical
  theories with a view to giving as comprehensive
  an account of computing practice as is possible
• revolutionary recognise that computer science
  as classically conceived is inadequate as a
  foundation for practical computing and artificial
  intelligence, and that this inadequacy cannot be
  resolved without reappraising the whole issue of
  how mathematical models and activities in the
  world are connected

15
Scope of CS 319

• Will use data modelling principles for databases
  as the arena within which to explore these broad
  issues concerning the role of theory in computer
  science
• Complementary to other modules, e.g. that concern
• formal approaches to software development
• artificial intelligence
• robotics
• personal views beyond scope of CS319 developed
  in CS405 An Introduction to Empirical Modelling

16
The principal theme of CS319

• How do theory and computing practice relate with
  specific reference to databases?
• Key ideas
• The counterpart of the classical theory of
  computation in databases is Codd's relational
  theory.
• There is an excellent correspondence between
  relational theory and practical database
  application of a certain kind.
• Relational databases can be seen as a precursor
  for two principal kinds of computer application
  environments for end-user programming and
  computer-based models of real-world state.

17
The principal theme of CS319

• How do theory and computing practice relate with
  specific reference to databases?
• Key ideas (cont.)
• There are aspects to the relational theory
  (especially those concerned with database design)
  that challenge the classical view of how an
  abstract mathematical model represents a
  situation in the world.
• Modern database applications oblige us to
  question the adequacy of relational theory, and
  to consider the extent to which a relational
  framework can cope with notions such as objects,
  interfaces and agents.

18
Content of CS319

• To put the key ideas in perspective, address the
  following topics
• An extended account of relational theory and its
  applications in respect of the development of
  relational query language and database design.
• The qualities of the relational model both as it
  was first conceived by Codd and as it has (and
  has not) been embodied in practical relational
  databases.

19
Content of CS319

•  
• A critical assessment of the relational database
  model from the proponents of alternative kinds of
  database, such as object-oriented and extended
  relational models.
• The possible limitations of the relational model
  where modern computing applications and
  technologies are concerned.
• An evaluation of the various database model
  critiques and a personal perspective on the
  future of databases.

20
Aspects of a Computer Science education

• Three kinds of activity involved in learning CS
• learning theory
• gaining practical knowledge and skills of
  computing
• developing a critical appreciation of how theory
  and practice are related
• Principal emphasis in CS319 is on the first and
  third aspects

21
Special lecture arrangements weeks 18 and 19

• Dr Hugh Darwen was an IBM consultant on
  databases
• he is our guest lecturer in weeks 18 and 19
• NOTE the lectures in week 18 will be on The
  Relational Model and SQL thus
• Monday February 21st at 2pm and 5pm in CS1.04
• Thursday February 24th at 2pm and 5pm in CS1.04
• the lectures in week 19 will be on
• Temporal Data and the Relational Model thus
• Monday February 28th at 2pm and 5pm in CS1.04
• Thursday March 3rd at 2pm and 5pm in CS1.04

22
Some context for the Theory of Databases

• Setting the scene
• Industrial perspectives
• Academic perspectives
• Sources for the CS319 module

23
Setting the scene for the theory of databases -
relations

• M J Ridley, University of Bradford (TLAD WS 2003)
• DB systems or DB theory or Why dont you teach
  ORACLE
• The modern age of databases may NB! - WMB have
  started with relational databases which despite
  their ubiquitous nature we should remember were
  looked an as of only academic interest at one
  time. What surely marks this era the early
  1970s out is the sound theoretical basis of that
  database model and the progress that was made
  because of this.
• But what age are we in now?

24
Setting the scene for the theory of databases -
objects

• Object oriented DBs emerged in the 1980s in
  response to the feeling that relational databases
  were inadequate for certain classes of
  applications ...
• Disadvantages of OODBs
• Brown, A.W. OODBs Applications in S/W
  Engineering, 1991
• no formal semantics
• loss of relational simplicity
• navigational queries
• no general query language

25
Setting the scene for the theory of databases
objrel

• Object oriented DBs emerged in the 1980s in
  response to the feeling that relational databases
  were inadequate for certain classes of
  applications ... The extent to which
  object-oriented and relational DBs can be
  unified, as in object-relational DBs, remains
  controversial
• Ridley cites Maier (2002) on object-relational
  systems
• David Maier speaks out, SIGMOD Record, 31(4),
  2002 http//www.acm.org/sigmod/record/issues/0212/
  
• theres not a good design theory the way there
  was with normal form relations for relational
  databases
• the new features at least double the
  number of design choices you have in every
  instance

26
Setting the scene for the theory of databases
the Web

• Ridley observes The Web has made many more
  people database users and heightened the
  awareness of databases generally it raises the
  issues of the semantic mismatch between the set
  oriented nature of SQL and most programming
  languages. cites
• The Asilomar report on Database Research, SIGMOD
  Record, 27(4), 1998 http//www.acm.org/sigmod/reco
  rd/issues/9812/
• Theorem The WEB changes everything.
• Corollary XML is the means
• and raises the question
• But where then is our dominant relational
  paradigm left?

27
Perspectives from the computing industry 1

• From Whatever happened to OODBs?
• http//www.leavcom/db_08_00.html
• In 1998 industry experts touted OODBs as a
  technology on the rise, well suited for the
  emerging Internet age OODBs would soon become
  the primary DB technology, supplanting RDBMSs,
  which were not designed to handle multimedia data
  Fast forward to today, and RDBs are still by
  far the most widely used DBs meanwhile,
  object-relational DBMSs have added object
  capabilities to relational DBs they are gaining
  popularity and are expected to outsell even RDBs
  by 2003

28
Perspectives from the computing industry 2

• Rick Catell (SUN Microsystems) Object-oriented
  DBs are doing just fine, and the news of their
  demise is greatly exaggerated. While their market
  isnt as big, they continue to be used in areas
  like CAD (computer-aided design) and
  telecommunications, where RDBMSs are not well
  suited.
• Michael Stonebraker (Informix) ODBMSs occupy a
  small niche market that has no broad appeal. The
  technology is in semi-rigor mortis, and ORDBMSs
  will corner the market within five years.
• quotations dating from around the year 2000

29
Perspectives from the computing industry 3

• International Data Corporation Opinion
• from OODB vs ORDBs, Steve McClure, 1997
• Because of the relative size of the RDBMS
  vendors marketing infrastructures, the ORDBMS
  market will surpass the size of the ODBMS market
  in the next five years. Nevertheless, putting
  object-oriented extensions on RDBMSs is
  tantamount to adding stereo radios and global
  navigation systems to horse-drawn carriages. You
  will have interesting enhancements, but the wrong
  base vehicle. In the end it wont be an
  appropriate vehicle for the information
  superhighway.

30
Perspectives from academia 1

• M J Ridley, University of Bradford (TLAD WS 2003)
• DB systems or DB theory or Why dont you teach
  ORACLE
• Practice without Theory is blind, Theory without
  Practice is sterile (Karl Marx)
• most people would agree that we need a
  theoretical framework in which to talk about DBs
  if the result is to be productive The questions
  that seem to me to need answering are What is
  that framework?

31
Perspectives from academia 2

• M J Ridley, University of Bradford (TLAD WS 2003)
• DB systems or DB theory or Why dont you teach
  ORACLE
• most people would agree that we need a
  theoretical framework in which to talk about DBs
  if the result is to be productive The questions
  that seem to me to need answering are What is
  that framework?
• Is it the simple relational model? How
  important is that framework? (In practical,
  educational terms how much of the syllabus is
  this?) Are there competing theoretical frameworks
  that need considering?

32
Perspectives from academia 3

• M J Ridley, University of Bradford (TLAD WS 2003)
• DB systems or DB theory or Why dont you teach
  ORACLE
• most people would agree that we need a
  theoretical framework in which to talk about DBs
  if the result is to be productive
• But is there still a firm theoretical foundation
  for what we teach? Does that matter? Are we in
  danger of training students in the features of a
  particular system (and a convenient web interface
  for it) rather than educating them about DBs in
  the broader sense?

33
Perspectives from academia 4

• M J Ridley, University of Bradford (TLAD WS 2003)
• DB systems or DB theory or Why dont you teach
  ORACLE
• What is the relationship of Database teaching to
  other aspects of the curriculum such as the study
  of data structures and object-orientation? And
  are there theories of these subjects?
• is there are more general object theory that
  informs programming and software development
  generally. If so can relational theory be located
  as a subset of that mainstream or does database
  stand out as an exception?

34
Perspectives from academia 5

• M J Ridley, University of Bradford (TLAD WS 2003)
• DB systems or DB theory or Why dont you teach
  ORACLE
• Ridleys own answer to these questions
• we are in danger of losing sight of firm
  theoretical ground especially as we are under
  pressure to include yet more additional
  facilities of DBMSs more so as object systems
  have failed to achieve commercial success
  commensurate with their early promise and the
  pressure of the web has shifted the focus of much
  database work.

35
Books on Classical Relational Databases

• Date, C.J. A Database Primer, Addison-Wesley
• Date, C.J. An Introduction to Database Systems,
  Addison-Wesley 8th edition, 2003
• Ullman, J.D. Principles of Database Systems 2nd
  edition, Computer Science Press 1982
• Ullman, J.D. Principles of Database
  Knowledge-base Systems Vols 1 2, Computer
  Science Press 1988
• Ullman, J.D. A First Course in Database Systems,
  2nd ed Prentice-Hall, 2002
• see http//www-db.stanford.edu/ullman/ullman-bo
  oks.html

36
Books on more general database issues

• Korth, H.F. and Silberschatz Database System
  Concepts, McGraw-Hill 2nd ed. 1991, 4th ed. 2001
• Brown, A.W. Object-oriented Databases
  Applications in Software Engineering, McGraw-Hill
  International Series in Engineering, 1991
• Gray, P.D. Logic, Algebra Databases,
  Ellis-Horwood Series, Computers and their
  Applications 29
• Kent, William Data and Reality basic assumptions
  in data processing reconsidered, North-Holland
  1987

37
Books on Relational Theory in its application

• Date, C.J. Relational Database Writings 1985-1989
  Addison-Wesley and other collections in this
  series
• Hugh Darwen and C.J.Date Foundations for
  Object/Relational Databases The Third Manifesto
  1998 Harlow, Addison-Wesley Longman
• C.J.Date, Hugh Darwen, M.A.Lorentzos Temporal
  Data and the Relational Model, Morgan-Kaufmann,
  2002
• (summarised in Chap. 23 of Dates Intro to DBs,
  8th ed)
• Dr Hugh Darwen was an IBM consultant on database
• until his retirement in 2004 he is our guest
  lecturer in
• weeks 18 and 19

38
Special lecture arrangements weeks 18 and 19

• Dr Hugh Darwen was an IBM consultant on
  databases
• he is our guest lecturer in weeks 18 and 19
• NOTE the lectures in week 18 will be on The
  Relational Model and SQL thus
• Monday February 21st at 2pm and 5pm in CS1.04
• Thursday February 24th at 2pm and 5pm in CS1.04
• the lectures in week 19 will be on
• Temporal Data and the Relational Model thus
• Monday February 28th at 2pm and 5pm in CS1.04
• Thursday March 3rd at 2pm and 5pm in CS1.04

39
Questions regarding status of DB theory 1

• Issues about the context in which DB theory
  applies
• To what extent do theories about databases in
  general apply exclusively to relational
  databases?
• To what extent is DB theory specific to
  applications?
• - technology / kind of activity / nature of
  environment
• it applies to the 1970s business context does
  it apply now?

40
Questions regarding status of DB theory 2

• Issues about the scope and nature of DB theory
  itself
• How does DB theory relate to objectives such as
• - supporting end-user programming?
• - representing real-world state?
• How does DB theory relate to other conceptual
  frameworks (e.g. data modelling in OO / UML / ER)
• to address broader issues (e.g. typing /
  non-procedural vs procedural models of
  computation)?