Toolbased Language Development

1
Toolbased Language Development

• J. Fischer, E. Holz, A. Prinz, M. Scheidgen

2
Outline

• Problems of modern software engineering The
  current situation
• SMILE-Project
• Role of common concepts
• A Toolbased language development Application

3
Situation
4
SMILE Project Overview

• SMILE Semantic Meta-model-based Integrated
  Language Environment
• Semantic
• Explicit representation of all semantic
  (meaning-giving) information
• Meta-model-based
• Model and Meta-model handling are central
• Integrated
• Uniform information handling independent of the
  level of abstraction, meta-modelling and
  executeability
• Language
• Focuses on language handling integration and
  provision of tools
• Support of small domain specific languages and of
  proprietary language extensions
• Environment
• Complete environment allowing all important
  operations

5
SMILE Vision

1. Define language semantics explicitly
2. Automatically generate tools from all
   descriptions
3. Increase the level of abstraction
4. Ease language development by providing tools
5. Handle all common sorts of descriptions
   grammars, graphics and semantics
6. Model and meta-model handling is in the centre
7. Main focus is on language handling and integration

7
6,7
2,3,4,5
3,6
3,6
2,3,4
2,5
1
6
SMILE Tool - Environment
Models
Grammar
Graphical representation
Static semantics
Transfor- mations
Metamodel
Metamodel
Semantics
SMILE
Generated tools
SMILE CORE
SMILE CORE
Translator
Checker
Parser
Editor
7
Example - Grammar Tool
metamodel generation
Metamodel
Grammar
processname type parameter ...
Agent
AgentType
Semantics Grammar to metamodel mapping
Parser
SYSTEM daemonGame signal newGame, endGame,
score(Integer) CHANNEL ser NODELAY
FROM ENV TO daemong WITH newGame, endGame, probe,
result FROM dae TO
gameId, win, lose, score ENDCHANNEL
processname ... metamodel
AgentClass.create(1) ...
8
Modular Language Definitions Common Concepts

• Key observation
• Despite syntactical differences languages
  usually share a set of common concepts!
• A modellbased language definition should exploit
  this for
• Language integration
• common concepts directly relate languages to each
  other
• common concepts support faster integration of new
  languages as well as the development of profiles
  and domain specific languages
• Reuse
• different languages have more commonalities than
  differences
• classification leads to hierarchies of concepts
• these hierarchies can be extended into semantic
  descriptions and tool development

9
Attributes of Common Language Concepts

• Common
• Different Languages can share the same concepts

• Abstract/Coherent
• The most abstract concepts should only describe
  single distinct characteristics

classifier

• Loose/Pluggable
• One must be able to use concepts idependently
  from other concepts

10
Development of a Set of Common Concepts

• Where to look
• Known elements What type systems, computational
  models, structure mechanisms do we know
• Decomposition Find abstractions
• Comparison What concepts do languages share and
  how can they modelled reusable

Set of common concepts
11
A Toolbased Language Development Application
12
You learned about

• SMILE
• Metamodel centred, model driven
• Automatic tool generation from semantic
  information
• Integrated language development via common
  concepts
• Common concepts Object orientation in language
  development
• Application A integrated SDL/UML Tool environment

13

• Thank You