Lessons learnt from adding unmaintained apps

1

Our experience with filetea and coquelicot illustrates the need for more careful consideration of the long-term viability of the apps we add into FreedomBox. Both coquelicot and filetea had no maintenance for 1 or 2 years when we picked them up.

Apps with active maintainers and a larger community should be preferred. Stable, feature-complete apps with low maintenance activity are an exception to this rule. However, in practice it might be hard to tell the difference between a feature-complete project and an unmaintained one (as in the case of coquelicot).

Coquelicot was also a disaster in terms of a user’s first impressions with FreedomBox. It shows up as the first app in alphabetical order in Buster stable (19.1). Users installed it and saw that it doesn’t work. Apple app store vets their apps thoroughly since a non-functional app might be seen by the user as a fault of the platform. It is no different in our case.

We should come up with a checklist for long-term viability of apps and use it when choosing between different apps in a category.

2 Likes
2

I think some red flags for identifying projects that are not serious:

  • Has no activity in the past year (when it is not a “done” project).
  • Has a single developer and has no project site, separate documentation, etc.
  • Calls itself a hobby/fun project or an experiment.
  • Uses outdated language or framework (such as Python2).
  • Lots of piled up pull requests or merge requests.

Some good signs can be:

  • Frequent releases.
  • Existed for a long time (> 5 years).
  • Popular.
1 Like