Dry Spell
Work continues to ship the stable version of 1.15.
Last week I wrote a blog post that explores some of the reasons behind the recent lack of new stable releases and patches. Summarizing, during the three-month stable release cycles there was always a certain amount of pressure on us to deliver the planned set of features. With Doomsday being a hobby project, eventually this generated enough negative energy that we had to reconsider some of our practices.
In plain language, we can't promise to deliver features on a specific date when there is no guarantee that time or energy is available for working on the project. While implicitly this has always been the intention, our release process didn't fully support it.
Thus we've hit the reset button and are reconsidering how releases will be made in the future. The appropriate model should be such that we can work on things that motivate us — this keeps the project moving forward — while bug fixes and other patches still get released regularly.
In practice, we are planning to use the autobuilder to make monthly patch builds for the stable version (when fixes are available). The increased granularity of stable versions also allows releasing code in smaller chunks, so that the stable branch isn't compromised by huge amounts of new code at once. Finally, further automating the stable releasing will improve our motivation to put effort into bug fixing since patches will be made available effortlessly — making patches is somewhat of a chore after the stable branch has had time to diverge from the latest work.
After 1.15.0 has been dealt with, expect us to make fewer promises about what will be included in the next release. New features will be released when they're done. However, there will be a larger number of small stable (patch) releases.
Last week I wrote a blog post that explores some of the reasons behind the recent lack of new stable releases and patches. Summarizing, during the three-month stable release cycles there was always a certain amount of pressure on us to deliver the planned set of features. With Doomsday being a hobby project, eventually this generated enough negative energy that we had to reconsider some of our practices.
In plain language, we can't promise to deliver features on a specific date when there is no guarantee that time or energy is available for working on the project. While implicitly this has always been the intention, our release process didn't fully support it.
Thus we've hit the reset button and are reconsidering how releases will be made in the future. The appropriate model should be such that we can work on things that motivate us — this keeps the project moving forward — while bug fixes and other patches still get released regularly.
In practice, we are planning to use the autobuilder to make monthly patch builds for the stable version (when fixes are available). The increased granularity of stable versions also allows releasing code in smaller chunks, so that the stable branch isn't compromised by huge amounts of new code at once. Finally, further automating the stable releasing will improve our motivation to put effort into bug fixing since patches will be made available effortlessly — making patches is somewhat of a chore after the stable branch has had time to diverge from the latest work.
After 1.15.0 has been dealt with, expect us to make fewer promises about what will be included in the next release. New features will be released when they're done. However, there will be a larger number of small stable (patch) releases.
Comments
Hopefully the next release will be the most bug free yet so I can get back to playing Hexen.
Would it help if we sent you cake?
We are currently in the final stages of preparing for the first 1.15 release candidate. Watch this space.
Still happy to send you cake in gratitude. Or brownies.
For some reason, he didn't appreciate it.
I'm sure I speak for the many other lurkers out there who use Doomsday and continue to eagerly look forward to the next release.
Keep up the great work!