[Qgis-developer] [Fwd: Re: [Qgis-user] Poll on website]

[Maxim Dubinin]

ML> Yes developing in branches helps keeps trunk more or less release
ML> ready but introducing new features inevitably will introduce new bugs
ML> / crash points. Although we have tried to introduce a culture of unit
ML> testing, this as not reallly taken hold, and that means we put the
ML> burden of testing and finding those bugs and crash points on our
ML> users. If we are going to do that it seems to me only fair that we
ML> indicate this with the 'unstable' moniker. There is a good quote that
ML> says 'software without unit tests is broken by design'....

Does it make 1.0.x broken by design too? I'm pretty
sure that "reliable" industry giants like ArcGIS should have unstable in their titles too, they
crush all the time. I don't think this is unfair to remove unstable
and have a "this software may/does contain bugs, use at your own risk"
mantra all over the website and qgis itself. This is reality, whatever
we call it.

Maxim