Affects Version/s: None
Fix Version/s: None
The concept of playgrounds, as we use it now, seems to be to provide a service to some "VIP" users, allowing them to get their code part of the automatic refactorings and thus be up-to-date with the development version of MATSim.
It is, however, more of a nuisance than an help. From the point of view of the user:
- Murphy's law seems to hold on the playgrounds as well, and "the big refactoring that unexpectedly breaks all my results" seems to happen more often close to a deadline than random. The solution becomes to work on a local branch until the paper is published and do the refactorings by hand once one have time, which defeat the "service" described above
- Playgrounds are often used to archive the code used for papers/project reports. Automatic refactorings on this code tend to break it, defeating this purpose as well
From the point of view of the core developer:
- Lots of time is lost refactoring unused code (because people are gone, because it is there only for archiving purpose, etc)
- The tremendous amount of commits in playgrounds makes the history of the core difficult to read
This seems to outweigh the benefits of playgrounds, in particular now that there is no dramatic changes in the API anymore. This would argue for letting each group/person create its/his/her own github repository and depend on MATSIM release or snapshots the Maven way.