dsls: the good, the bad, and the ugly tihamér levendovszky institute for software-integrated...

5
DSLs: The Good, the Bad, and the Ugly Tihamér Levendovszky Institute for Software-Integrated Systems Vanderbilt University Nashville, TN

Upload: piers-atkins

Post on 04-Jan-2016

215 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: DSLs: The Good, the Bad, and the Ugly Tihamér Levendovszky Institute for Software-Integrated Systems Vanderbilt University Nashville, TN

DSLs: The Good, the Bad, and the Ugly

Tihamér LevendovszkyInstitute for Software-Integrated Systems

Vanderbilt UniversityNashville, TN

Page 2: DSLs: The Good, the Bad, and the Ugly Tihamér Levendovszky Institute for Software-Integrated Systems Vanderbilt University Nashville, TN

Background

• Related research experience– Member of the Generic Modeling Environment

development team (VU)– Leader of the Visual Modeling and Transformation

System team (BUTE)• Related industrial experience– Architect, Designer, Consulting (Hungary)– Project manager: Mobile Innovation Centre,

Hungary

Page 3: DSLs: The Good, the Bad, and the Ugly Tihamér Levendovszky Institute for Software-Integrated Systems Vanderbilt University Nashville, TN

UML Class DiagramUML Statechart DiagramUML Use Case DiagramVMTS Resource ModelVMTS Database ModelVMTS Control Flow ModelVMTS Rule Editor Model

Page 4: DSLs: The Good, the Bad, and the Ugly Tihamér Levendovszky Institute for Software-Integrated Systems Vanderbilt University Nashville, TN

„The Good”

• Potentially increase the productivity• Self-evident to express problems and solutions• Higher abstraction level• More or less mature: the language

engineering knowledge is there – applicable to industrial projects

• Part of a tool chain to underpin generative techniques: real reuse support

Page 5: DSLs: The Good, the Bad, and the Ugly Tihamér Levendovszky Institute for Software-Integrated Systems Vanderbilt University Nashville, TN

„The Bad, and the Ugly”

• Proprietary languages, the knowledge has not been transfered yet to tools– only part of it in several tools

• Takes time for the industry– To understand it

• Generative technique often misused• Thought to solve all software problems

– To take the risks• Introducing and estimate a new technology and tool• Paying the learning curve

– To think in more then one project• A DSL with a generator pays off if used more than once• Needs predicting the future projects

• Composition (integration) and evolution is unsolved yet