Application Generators

1
Application Generators

• Advanced Seminars in Software Reuse
• Leandro Marques do Nascimento

2

• Every minute spent on infrastructure programming
  is a wasted minute.
• Juval Lowy, .NET Software Legend, April 2003

Code generation is a formalized version of
copy-and-paste programming. (CGN, 2005)
3
Software Reuse
4
(McIlroy, 1968)

• Mass Produced Software Components
• Felt that component libraries could be
  effectively used for numerical computation, I/0
  conversion, text processing, and dynamic storage
  allocation.

5
(Freeman, 1983)

• Not only source code can be reused
• Different types of artifacts which are not
  limited to source code fragments can be reused
  including design structures, module-level
  implementation structures, specifications,
  documentation, transformations, and so on.

6
(Biggerstaff and Perlis, 1989a) (Biggerstaff and
Perlis, 1989b)

• Commonality among the software reuse techniques
  used
• Abstraction essencial feature in any reuse
  activity
• Selection selects the appropriate artifacts to
  the specific domain.
• Specialization specialize the artifacts through
  transformations, constraints, parameters or other
  refinement
• Integration combine a collection of selected and
  specialized artifacts into a complete software
  system

7
(Krueger, 1992)

• Software reuse is the process of creating
  software systems from existing software rather
  than building software systems from scratch.
• Application generators operate like programming
  language compilersinput specifications are
  automatically translated into executable
  programs.
• Using Generators the code expansion
  (input/output) may reach 2000 lines of code per
  day.

System specification
Generators
what to do
System implementation
how to do
8
(Krueger, 1992)

• Abstraction

System Architecture (Domain Dependent)
Textual Specifcation Languages (4th generation
languages)
Graphical Diagrams
Templates
Interactive menu driven dialogs
Structure-oriented interface
9
(Krueger, 1992)

• Selection
• Depends on domain covarage
• Application generators are often highly
  specialized, limiting their domain coverage
• Specialization
• Primary task of a software developer, by
  providing an input specification
• Conventional generators business-oriented,
  data intensive applications. Report generation,
  data processing and displaying
• Expert generators expert knowledge to solve
  problems in a particular domain
• Parser and compilers generators the most known
  examples of application generators, such as Lex
  (Lesk and Schmidt, 1979) and Yacc (Johnson, 1979)
• Integration
• Generators that produce a complete executable
  system do not need integration

10
(Krueger, 1992)

• Software life-cycle using application generators
  (ideally)

Automated
Automated
System specification (requirements)
Architectural design
Detailed design
Coding
Testing
Automated
Unnecessary
11
(Sametinger, 1997)

• Reuse techniques
• Compositional Reuse mounting systems from
  unmodified components (ideally)

Component repositories
Component composition
Code and design Scavenging
Code
Design
Blocks of code copied from other system
Large blocks of code. Details deleted. Design
retained.
12
(Sametinger, 1997)

• Reuse techniques
• Generative Reuse reuse of a generation process
  instead of reuse of components. Examples lexical
  analysers, parsers and compilers.

Application Generators
Transformation Systems
Language based generators
Programming Languages
13
(Sametinger, 1997)

• Generation vs. Composition

14
(Czarnecki and Eisenecker, 2000)

• Domain Engineering is the activity of
  collecting, organizing, and storing past
  experience in building systems or parts of
  systems in a particular domain in the form of
  reusable assets
• Vertical X Horizontal Domains
• It encompasses

15
(Czarnecki and Eisenecker, 2000)
16
(Czarnecki and Eisenecker, 2000)

• Generative Programming (GP) is about automating
  of intermediate and end-products (i.e.,
  components and applications)
• Domain Engineering helps in GP

System Families
Model Problem Space
Find Components
Model the Config. Knowledge
17
(Czarnecki and Eisenecker, 2000)

• Steps in GP

Domain Scoping
Feature and Concept Modeling
Design Common Architec.
Identify Components
Specify DSLs
Specify Config. Knowledge
Implement Components
Implement DSLs
Implement Using Generators
18
Application Generators
19
(Neighbors, 1980)

• DRACO
• First system to support the transformation of
  high-level domain-specific programs to executable
  code
• DE based
• Idea organize software construction knowledge
  into a number of related domains
• DRACO domain encapsulates

DSLs
Optimization Transformations
Transformational Components
Domain-Specific Procedures
Transformation Strategies
Parse Tree
Rules
Refinements
Domain Dependant
When a set of transformations can be performed
algorithmically
20
(Neighbors, 1980)
DRACO Practical Example
21
(Batory, 1996)

• GenVoca
• ...building software generators based on
  composing OO layers of abstraction, whereby
  layers are stacked.
• Roughly corresponds to the refinement occurring
  in OO frameworks
• Steps

Identify layers of abstraction
Put more specialized layers on top
Layer below as parameter to the above
Families of parameterized layers
22
(Batory, 1996)

• GenVoca Model (layering with Booch C Data
  sctructure)

23
(CGN, 2005)

• Four key benefits of code generation
• Quality increased because of the fixes in
  templates applied to all code base
• Consistency generated APIs are consistent in
  class structure
• Productivity generators use a fraction of time
  of an engineer. So, the engineer can concentrate
  on more creative solutions
• Abstraction layer of abstraction between design
  and code base. Definitions input files hold the
  business schema in a abstracted declarative form
  which can be inspected by domain experts

24
(CGN, 2005)

• Code generators models
• Code Munger
• Inline Code Expander
• Mixed Code Generator
• Partial Class Generator
• Tier Generator
• Domain Language

25
(CGN, 2005)

• Code Munger Example JavaDOC

• Q accurate with the code
• C consistent and reliable

• P no external docs
• A little abstraction

26
(CGN, 2005)

• Inline Code Expander Examples CSQL, JSQL

• Q infra-structure code reduced
• C consistent code adaptable to other DBs

• P save time in DB intensive applications
• A easier inspections in all SQL commands

27
(CGN, 2005)

• Mixed Code Generator Examples CSQL, JSQL

• Q infra-structure code reduced. Better with
  IDEs
• C consistent code adaptable to other DBs
  according to templates

• P save time to focus on SQL
• A easily extracted from code

28
(CGN, 2005)

• Partial Class Generator Example Visual Studio
  .NET

• Q database layer controlled by templates. OO to
  DB errors reduced
• C consistent in all generated classes and API

• P save time to upper layers and not worrying
  with persistence bugs
• A abstracted into definition files and templates

29
(CGN, 2005)

• Tier Generator Example Hibernate

• Q equivalent to quality in templates
• C class structure reliable

• P entire tier generated reduces much time
• A tier abstraction is held outside of the code

30
(CGN, 2005)

• Domain Language Examples Mathematica, Matlab
• Types, syntax and operations mapped directly to
  concepts in the domain
• Easier to represent specific structures that
  general purpose languages do not provide, such as
  multi-dimensional matrix
• Other examples
• COBOL COmmon Business Oriented Language
• Domain-Specific Language (DSL) Tools for
  Microsoft Visual Studio 2005
• CSound to create audio files
• GraphViz define graphs and get a visual
  representation

31
(CGN, 2005)
32
Qualiti Coder (Qualiti, 2005)

• Java e C
• Eclipse Plugin
• Wizard based
• Layer architecture Presentation, Data and
  Communication (PDC)

33
(Velocity, 2005)

• Based on Model-View-Controller
• Allows to separate Java programmers from web
  designers
• Can be used to generate SQL, PostScript and XML
  from templates
• Also can be used to generate source code and
  reports

34
(Velocity, 2005)

• GCodCESAR Usa o Velocity

35
(J2ME Polish, 2005)

• J2ME GUI Code generation
• Based on CSS (polish.css file)
• Bitmap fonts

36
(J2ME Polish, 2005)

• The tool makes use of preprocessing directives

37
Conclusions

• Difficulties to be covered
• Domain peculiarities
• Code Size and efficiency
• Libraries for algorithmic problems

• Future to generators
• More readable languages (5th generation)
• Large domain coverage
• Efficient Drag-and-Drop generation

38
References

• (MacIlroy, 1968) MacIlroy, M. D. Mass Produced
  Software Components. Nato Software Engineering
  Conference. 1968.
• (Freeman, 1983) Freeman, P. Reusable software
  engineering Concepts and research directions. In
  Workshop on Reusability in Programming (Newport,
  R. I., Sept.). ITT Programming, Stratford, Corm.,
  pp. 2-16.
• (Biggerstaff and Ritcher, 1987) Biggerstaff, T.
  J. Ritcher, C. Reusability framework,
  assessment, and directions. IEEE Software. 4, 2
  (Mar.), 4149. 1987.
• (Biggerstaff and Perlis, 1989a) Biggerstaff, T.
  J. Perlis, A. J. Frontier Series So filoare
  Reusability Volume IConcepts and Models. ACM
  press, New York. EDS. 1989.
• (Biggerstaff and Perlis, 1989b) Biggerstaff, T.
  J. Perlis, A. J. Frontier Series Software
  Reusability Volume IIApplications and
  Experience. ACM Press, New York. EDS. 1989.
• (Qualiti, 2005) Qualiti Coder. Available on
  http//coder.qualiti.com, accessed in November,
  2005.
• (Sametinger, 1997) Sametinger, J. Software
  Engineering with Reusable Components.
  Springer-Verlag, 1997.
• (Krueger, 1992) Krueger, C. W. Software Reuse.
  ACM Computing Surveys, Vol. 24, No. 2, June 1992
• (Lesk and Schmidt, 1979) Lesk, M. E. Schmidt, E.
  Lex A Lexical Analyzer Generator in the UNIX
  Programmers ManualSupplementary Documents. 7th
  ed. AT T Bell Laboratories, Indianapolis, Ind.
  1979.
• (Johnson, 1979) Johnson, S. C. Yacc Yet Another
  Computer-Compiler in the UNIX Programmers
  ManualSupplementary Documents. 7th ed. ATT Bell
  Laboratories, Indianapolis, Ind. 1979.
• (Czarnecki and Eisenecker, 2000) Czarnecki, K.
  Eisenecker, U. W. Generative Programming
  Methods, Tools, and Applications. Addison-Wesley,
  May, 2000.

39
References

• (Czarnecki and Eisenecker, 1999) Czarnecki, K.
  Eisenecker, U. W. Components and generative
  programming. In Proceedings of the 7th European
  software engineering conference held jointly with
  the 7th ACM SIGSOFT international symposium on
  Foundations of software engineering ESEC/FSE-7.
  vol. 24. October, 1999.
• (CGN, 2005) Code Generation Network. Available on
  http//www.codegeneration.net/, accessed in
  November, 2005.
• (Pulvermüller and Speck, 2000) Pulvermüller, E.
  Speck, A. Towards generative components. In ACM
  SIGSOFT Software Engineering Notes archive. Vol.
  25 Issue 2. pp. 22-24. ACM Press, March, 2000.
• (J2ME Polish, 2005) J2ME Polish. Available on
  http//www.j2mepolish.org/, accessed in November,
  2005.
• (Velocity, 2005) Jakarta Velocity. Available on
  http//jakarta.apache.org/velocity/, accessed in
  November, 2005.
• (Neighbors, 1980) Neighbors, J. M. Software
  construction using components. Ph. D. Thesis,
  Department Information and Computer Science,
  University of California, Irvine, 1980.
• (Neighbors, 1989) Neighbors, J. M. Draco a
  method for engineering reusable software systems.
  In Source Software reusability vol. 1, concepts
  and models. pp. 295-319. ACM Press, 1989.
• (Batory, 1996) Batory, D. Software System
  Generators, Architectures, and Reuse. In Fourth
  International Conference on Software Reuse.
  Tutorial Notes, Orlando, 1996.

40
Application Generators

• Advanced Seminars in Software Reuse
• Leandro Marques do Nascimento