I’ve been rushing DragonMail to something which I wouldn’t be embarrassed for other people to play, not just look at carefully selected screen shots. The goal was a soft alpha release this weekend, and the release platform is obviously all in place (you’re reading it), but:

I thought it would be sensible to do a complete playthrough (OK, it’s completely open-ended at the moment, but there’s a definite goal) on the release candidate (0.3.11) before announcing it, not least to see how good my guess was as to how long a run would take. That wasn’t too bad – about 90 minutes game play when I was aiming for just under an hour – but did it ever uncover a lot of problems, falling into roughly four categories:

• Things which could be improved that I hadn’t realised were deficient until running into them while properly playing
• Things which I know are bugs, but might be seen as potential improvements by others
• Known bugs that were previously seen as minor irritants having major impact on through-playability
• Completely new bugs

The unresolved bug list jumped from 5 to 19 in the course of 2 hours. 6 of which are so bad (such as the poor mouse that spawned buried beneath a temple floor, who I could only rescue by running the savefile in the Unity editor and using god-like powers) that there’s no way I’m prepared to announce a release until they’re fixed. And another 6 which are hopefully fairly quick-win improvements, which I’m determined to include in that first official release. So even next weekend might optimistic.

Lesson (which you’d think I’d been in the programming business long enough to have learned): don’t just allow time for testing, allow time for fixing the problems that testing will reveal.