Escrúpulos Una piedra en la conciencia del software libre, calidad y la vida misma

After releasing 3.95, I started a discussion about the schedule for future releases. There were two issues, I think, which affected the possible schedules. First, the feasibility to make quick releases and build upon that to have a stable release in a sensible timeframe. Second, whether we should move to a tighter integration with KDE SC or not.

My aim is to have a stable release within 6-8 months. For that, we had to change our usual 5-6 month release cycle, and have quick releases which would allow us to build upon the feedback from battlefield.

This first release surfaced some problems in different distros that we didn't quite expect, like having built some stuff based on libs which are still moving targets, or bugs in our build scripts, which were a little weak due to our inexperience with cmake. We didn't fully use cmake until porting to KDE4, and then this is a big application, so some problems didn't really surface until tested in a much broader spectrum of environments than those of the development team.

With all this in mind, we decided to have at least 3 quick release cycles. Each cycle will be 4 weeks of full development, 2 weeks for stabilization and string-freeze, release, rinse and repeat. Our next string-freeze starts on February 1st, followed by the release of version 3.96 on February 15th. I think this fits nicely with the KDE SC 4.4 schedule, to allow the translators some time when they are not focused on the main release. At least I hope it works that way.

The next releases will be about March 30th for 3.97 and May 15th for version 3.98, hopefully the latter will be Release Candidate 1.

As for the second issue, whether we should integrate further with KDE SC or not, if things are ok at that point, we intend to move KMyMoney to kde-review, for inclusion in KDE SC 4.5, and release as part of extragear in August (or July, I have to check the schedule in detail). If we don't make it, we'll try again for SC 4.6.

Of course, no plan survives the first engagement, so we'll adjust based upon the experience of these quick releases. The lessons so far have proven to be valuable and we have solved many errors in these last days. We are also working hard on migrating some of our custom widgets to MVC, so expect SVN to be a little bumpy until the end of month, when things will calm down as we move to stabilization mode.