Lessons learned from real DSL experiments

Over the years, our group, led by Bob Baker, designed and implemented three domain-specific languages for use by outside people in real situations. The first language described the communication format of messages used by NATO to specify command-and-control messages between people and equipment; the processor we generated checked these messages for consistency. The second language was in part graphical, part textual, and was used to demonstrate how naval ship formations were constituted and the constrained movements they could undergo. The last language was a mixture of graphics, text, and declarative information specified using three different COTS products. It was used to describe census survey "instruments," used to collect census data in the field. The code generated was to be installed in the survey takers\´ laptops. Each of these was actually a prototype for what would have taken more time to engineer and polish before putting into practice. Although each effort was essentially successful, none of the languages was ever followed up with the subsequent engineering efforts that we expected or at least hoped for. The first two were demonstrated and reviewed informally. The last effort was more seriously reviewed, in that training sessions and a formal review process were undertaken to evaluate the potential effectiveness of the product. Herein I elaborate where these language efforts succeeded and where they failed, gleaning lessons for others who take the somewhat risky step of committing to develop a DSL for a particular user community.