-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

> Apply for edit_bug status (IIRC it was on the login screen a while
> back). But then be very careful not to confirm duplicates, confirm bugs
> that are already fixed.

Similar thing was asked already - how ? It is not written anywhere and an 
"outsider" has no way to know it. I support the idea of the HOWTO mentioned 
in other threads, that could help a lot. 

> > since it is not possible to easily 'confirm' the bug, we "mark" almost
> > everything as 'non-important' and it never gets fixed, to the
>
> frustration of
>
> > the bug reporters, testers and, in the end, users.
>
> Well, this is better than having millions of duplicate and bad bug
> reports, as it was before, and the developers are so swamped they stop
> using bugzilla.

Buchan, but if we do not trust the bug reporters, then what's the point of 
having public Bugzilla ? The reports would go ignored one or the other way. 

I know that people file bad reports and misreport bugs. That's normal. But 
usually the approach taken is similar to something like below, instead of 
ignoring the reports : 

1) somebody files a bug
2) QA tests and tries to verify it
3) if the bug is there, mark as 'NEW' and/or 'ASSIGNED' (depends on the 
workflow in the company)
4) if the bug is invalid, gets marked as 'INVALID' or 'WORKSFORME'
5) developers fix assigned bugs, put them in 'FIXED' state
6) QA tests the fix and either 'CLOSE's the bug or reassigns it back

In our case, since Mandrakesoft's QA obviously is not able to do all this 
because of the sheer amount of bugs and various combos of the hardware and 
software installed, let's establish a network of trusted volunteers 
"prescreening" the bugs for each component. These people should be very 
responsive to the reports and Mandrakesoft would colaborate with them 
closely. Something similar works for Linux kernel, where there was the 
discussion, that "Linus does not scale" some time ago. They solved it by 
establishing several "patch penguins", who take bug reports and patches, test 
them and only then feed them upstream to Linus for inclusion. I think, that 
something similar could work here too. And if Mandrakesoft rewards them 
somehow (club points or a free powerpack or I do-not-know-what), it could 
work for both sides.


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQE+Z5byn11XseNj94gRAqIBAJ9qtGcuzn85MYCEqAPvAZhmsdkA4ACfaZwy
CX4P8Vm9gqrFEyozpY53Xew=
=vtbQ
-----END PGP SIGNATURE-----


Reply via email to