We should Stop Claiming Generality in our Domain-Specific Language Papers

TL;DR

This paper argues against claiming broad generality for domain-specific languages, emphasizing the benefits of specificity and practical co-development over overly general language claims.

Contribution

It challenges the common belief in the community, advocating for recognizing the value of specific, application-tailored domain-specific languages instead of overly general ones.

Findings

General DSLs lead to more boilerplate code

DSLs are co-developed with applications, making them inherently specific

Acceptance should focus on good DSL engineering practices

Abstract

Our community believes that new domain-specific languages should be as general as possible to increase their impact. However, I argue in this essay that we should stop claiming generality for new domain-specific languages. More general domain-specific languages induce more boilerplate code. Moreover, domain-specific languages are co-developed with their applications in practice, and tend to be specific for these applications. Thus, I argue we should stop claiming generality in favor of documenting how domain-specific language based software development is beneficial to the overall software development process. The acceptance criteria for scientific literature should make the same shift: accepting good domain-specific language engineering practice, instead of the next language to rule them all.