Les Mikesell wrote:
Ralf Corsepius wrote:
>>>>> Problem: We need more and wider testing. Why don't we get more
>>>>> testing?
>>>> .. because this work is not attractive. It's boring work without
>>>> proper credit in open source community.
>>> Right. Furthermore, testing implies finding bugs, discussing,
>>> struggling
>>> and arguing with package maintainers and upstreams. Not necessarily a
>>> way to make friends :)
>>>
>>> However, I think the primary cause in is Fedora's work-flow and
>>> Fedora's
>>> infrastructure. I find them not to be really helpful to such
>>> endeavors.
>> I do think Windows has improved a lot since they added the crash
>> reporter. OS X has one
>
> These are closed source OSes - They don't have any alternative but such
> "user participation programs" - OSS has alternatives.
Beg your pardon, but having the option (requirement?) to fix broken
stuff myself has never been all that appealing to me as an aspect of
open source. I look to it more for the benefits of re-using code that
is already well tested. Of course that doesn't work out all that well
in a project that keeps changing things...
>> - and I though Ubuntu included one too although I haven't seen
>> anything trip it.
> Gnome had one for many years (bug-buddy), ... I don't recall having seen
> it providing any substantial improvement to Gnome.
>
> Now the kernel also has one (kerneloops) ... We'll see if it will
> provide improvements.
How does that work? I'd think a dead kernel or one that doesn't boot
would have a hard time reporting it's problems.
A kernel oops is not a crash and it doesn't render the kernel
unbootable. See
http://en.wikipedia.org/wiki/Linux_kernel_oops .
> My expectations on such tools are very low. Many users switch
them off
> and developers/maintainers tend to ignore them as noise.
If the developers ignore the information, then at least they should
stop blaming the lack of testers for the lingering bugs.