GIMP developer Martin Nordholts has written an interesting article about the reasons for the delay of GIMP 2.8.

First off: GIMP is a project that lives from the free time of the developers – the friends & family situation changes all the time. Sometimes you’ve got time to hack on GIMP, sometimes not. The problem is that if a developer cannot finish a feature completely before being interruped by longer pauses, this means that there are incomplete features on the main branch. This situation delays the work of others, and in the worst case blocks whole releases.

The second point concerns the main branch itself – most work on features is done in this main branch. Many developers who join and contribute stuff are working directly in the main branch, because they want to get feedback for their stuff as soon as possible from the wide user base of GIMP – that’s the fun of it! However, it also blocks faster development cycles, because often the work is not finished.

Another thing that Martin mentions: sometimes there is a point where things get boring, for instance if you have to work on older things or run into problems. However this problem only appears if the feature someone is working on is a big one.

So the solution would be to split development into different branches. New features should be developed in their own branches and only be merged if they’re complete. Martin is currently working on a Web-based tool for a better scheduling for GIMP release cycles. More information to come.

You can read Martins full blog post on his blog: chromecode.com