Comparison between different abstraction level programming: Experiment definition and initial results

Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review

Abstract

Domain-specific languages and especially domain-specific modeling languages (DSML) are mentioned to achieve 5-10 times performance gains compared to traditional software development practices due to raising the level of abstraction. The data for the cases where these gains have been witnessed is usually not available. Therefore, in this paper, we introduce a simple but comprehensive and affordable experiment framework which can be utilized for measuring the benefits and drawbacks of DSMLs in an open fashion, i.e. publishing the data and the results and enabling a possibility to repeat the experiment by others. In this paper, we also present our own experiences and initial results about the benefits and drawbacks. We found the benefits of DSML to be clear if the applications have to be implemented daily and especially if the platform continues to evolve and the existing applications have to be updated to correspond to the changes. The benefits of utilizing DSML in a domain where the platform continues to evolve came as a surprise and needs further study.
Original languageEnglish
Title of host publicationProceedings
Subtitle of host publication7th OOPSLA Workshop on Domain-Specific Modeling 2007
PublisherAssociation for Computing Machinery ACM
ISBN (Print)978-1-59593-865-7
Publication statusPublished - 2007
MoE publication typeA4 Article in a conference publication
Event7th OOPSLA Workshop on Domain-Specific Modeling 2007 - Montreal, Quebec, Canada
Duration: 21 Oct 200725 Oct 2007

Conference

Conference7th OOPSLA Workshop on Domain-Specific Modeling 2007
CountryCanada
CityMontreal, Quebec
Period21/10/0725/10/07

Keywords

  • Domain-specific modeling languages

Fingerprint Dive into the research topics of 'Comparison between different abstraction level programming: Experiment definition and initial results'. Together they form a unique fingerprint.

  • Cite this