Posting it here, since ideas are in rather initial state :) Just my thoughts for today. Don't think it's about our thread problems, it's just about whole desktop. Btw, we can also improve stability. Process separation is great thing to make software more stable. Currently we are running single instance, and it's not a good thing for sure. It would be nice to switch to process-per-document model. So as a blog replacement :) Current GNOME quality assurance relies mostly on user testing. The typical flow involves bug reporter who met the problem and developer both involved in interaction to fix the issue. The number of bugs submitted currently and the time of response show that this approach is not so effective as it should be. Moreover it's not acceptable to rely on users in software testing just because users expect quality product avoiding testing role. There are modern techniques that allow developers to raise the quality of software without end-user involvement. They are proper high-level software models specification and testing, code coverage testing, static and dynamic analysis, formal specification and so on. There are even free tools developed, like architecture description languages, static code analysers like JInt, FindBugs and splint. There are test coverage generators and testing frameworks. Of course they often relies on high-level languages like Java and doesn't work so well with C, but they can be adopted to GNOME developement. So it would be interesting to know what do others think about this problem and the way we can move to better product quality.
Attachment:
signature.asc
Description: =?koi8-r?Q?=FC=D4=C1?= =?koi8-r?Q?_=DE=C1=D3=D4=D8?= =?koi8-r?Q?_=D3=CF=CF=C2=DD=C5=CE=C9=D1?= =?koi8-r?Q?_=D0=CF=C4=D0=C9=D3=C1=CE=C1?= =?koi8-r?Q?_=C3=C9=C6=D2=CF=D7=CF=CA?= =?koi8-r?Q?_=D0=CF=C4=D0=C9=D3=D8=C0?=