[Qgis-developer] Quality assurance

[Agustin Lobo]

I'm happy this subject is getting such a large attention.
I include the users lists in my message as this thread is most
interesting for users (actually, this is a users question
rather than a developers question).

First of all, I want to stress that nothing I've said implies that I
do not think
"that developers know about their responsibility". That's not the problem.
I also think that many errors are fixed very quickly, and if this were
not the case,
I would not complain: I would just not use QGIS.

My concern is not that QGIS has many bugs, but that users do not have a
clear list of the most important current problems they can find if they
use a given version of QGIS. This is why I think we need to create a team of
committed users that would maintain a page with the most important
operational problems
(and I obviously show up for becoming a member of that team).

The proposed list is not going to be very useful for developers, who
already have the "issues" system to which
some of us (users) try to contribute. But is very important for users
to know the limits
within which they can safely operate. The problems I had with the Join
or with the georeferencing were not
worrying on their own: both problems could be easily circumvented. The
problem is the loss
of confidence because the user thinks: "what other errors am I going
to find next?" or even worse: "might it
be that the processing I've done up to now was wrong?"

As  I said, QGIS is good, thus highlighting the most important
problems (according to the experience of users) is
actually reassuring.

A second question relates to the popular comment "you cannot expect
people to work for you all the time for free".
Users reporting errors work for free also, in the sense that
developers do not pay them. And, in OS, the rule is
that most users do not pay developers (and the reverse), which does
not mean that developers or users work for free.
In the case of R, many universities, research centers and hospitals
provide funding, mainly in the form of having
permanent technicians working on R development as a part of their job.
But this happened once R reached stability and
universities started to save on commercial licenses.
I think that the funding problem with QGIS is that it has not been
adopted yet by any university as its main GIS. And
I think the problem is that, at the moment, QGIS is still seen as too
buggy. A second reason is that many users do not
see any real advantage because the university provides campus
licenses. Well, in that case, QGIS must offer features that
other programs do not. And I personally find that a big one is the
link to other platforms such as Grass, Saga, OTB and R.

Agus

2011/12/19 Ivan Mincik <ivan.mincik at gista.sk>:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> The problem with testers could be that people could not have access to
> all supported platforms (for example I am able to test something only in
> Debian Squeeze). Most of testers do not have Macs, so it could be a
> problem to say that I will guarantee that this or that feature is OK,
> because I did test only for my platforms. So again, I am returning back
> with the question on test suite.
>
> By the way what are the Linux supported platforms (at least which Ubuntu
> versions)?
>
>
> At least, Big thanks to all devs for backporting fixes to maintenance
> releases.
>
> - --
> Ivan Mincik, Gista s.r.o.
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.10 (GNU/Linux)
>
> iEYEARECAAYFAk7vFvcACgkQVqso/9cUsCwxKACgxrJ3oLXdxikntRuQzut0c3h0
> Z8oAn0iyA+XhLWGClkolPTL/qMds8M+s
> =mzSF
> -----END PGP SIGNATURE-----
> _______________________________________________
> Qgis-developer mailing list
> Qgis-developer at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/qgis-developer