Re: Group up orphaned packages in order to find adopter faster
On Tue, 28 Oct 2008, Stefano Zacchiroli wrote:
Enrico Zini have ad for a while some code which find package
similarities, in fact I believe it is the one used on
packages.debian.org to show "similar packages".
Hmmmm, what means "similar packages"? I really can not imagine
some code that detects "similarities".
I suggest that instead (or maybe in addition) of doing your own
partitioning, you try to integrate with that code.
I might add the fact that I'm in the process of writing QA
tools for Debian Pure Blends (formerly known as Custom Debian
Disrtibutions) like Debian Med and Debian Edu which group
packages according to their use. I started with an overview
of those working fields which we are calling tasks and which
are implemented as metapackages. The overview pages are
http://debian-med.alioth.debian.org/tasks/ (Debian Med)
http://cdd.alioth.debian.org/science/tasks/ (Debian Science)
and I just have some raw code that produces an overview about the
bugs against packages grouped like that. This is *not* an
announcement yet, because these pages are lacking some features
I regard important but once you are talking about QA of "similar
packages" I really think that it fits the topic and you might
like to have a view on
The code is available at
and I plan to add more stuff like that. The idea to group packages
in task files according to the interest of specific user groups IMHO
leads to the feact that you might be able to point specialised users
to bugs they are interested in.