When I updated my tool to a new version, the whole tool entry was made invisible because it’s status was changed to “unpublished”, even though the old version passed the review. This could cause confusion about tools disappearing from the repo page.
I suggest that the old version should still appear on the tools page (with a download link to the older file, old description, old screenshots) while the new one is being revisited.
[edit]
In addition, I think that a manual review by a member of the Renoise team for every bugfix-release of a tool could quickly become too cumbersome, both for team members and authors alike. I updated my tool a few hours ago and somebody already found a bug. I don’t want to upload the bugfix-release to tools.renoise.com knowing that tomorrow night I might already upload the next release because I don’t want to get on the nerves of the team members. I see three possible solutions to this problem:
Option 1: The authors need to find a way to distribute “testing” releases independently. A problem with this option is the reduced visibility of the release (it’s not on tools.renoise.com) and scattering of the distribution channels of the individual tools and versions (the [newest version of] tool XYZ is not yet on tools.renoise.com, you need to download it from someplace.com) which defeats the purpose of having a centralized repository. The good thing about this is that it would force the authors into a slower, more stable release cycle.
Option 2: Add an option to download an unverified version to the repo, with a big fat warning preceding the download page, and forcing the user to sign in. Much like addons.mozilla.org handles this. The problems with this option are 1) more work for the Renoise team to implement this feature and 2) the danger of having a data-destroying, man-eating, wife-cheating plugin in the repo. The good thing about this is that users can always download bleeding-edge versions of their favorite tools from the repo, maintaining it’s centralized nature and authority as “the place to download add-ons”.
Option 3: Enable users to peer-review the tools. Every user would need to be able an unverified version of a tool and then vote for or against that tool. If a tool has a certain number of votes, it gets “published”. The downside with this solution is, in addition to the ones mentioned in option two, that this system would be vulnerable to malicious users who register accounts to place a vote.
This is a difficult problem. What do you suggest?
[/edit]
