[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Package status in Debian-med website



Hi,


On Thu, Nov 30, 2006 at 10:10:38AM +0900, Charles Plessy wrote:
> Le Wed, Nov 29, 2006 at 10:29:02PM +0100, Andreas Tille a écrit :
> > On Wed, 29 Nov 2006, Michael Hanke wrote:
> > >	dinifti - Debian package not available
> > >
> > >This should be updated as dinifti aka dicomnifti is in unstable.
> 
> > Perhaps we should write some kind of "codex" for developers of
> > Debian-Med related packages.  After they notice that one of
> > their packages hit testing they should make sure that two
> > people will be informed:
> > 
> >      1. The person who cares for Debian-Med web site
> > 
> >      2. The maintainer of the debian-med source package
>
> > Apropos up to date:  The latest Debian-Med news is older than one
> > year which makes not a really good impression. So we might be able
> > to add a news item featuring
> > 
> >    * Debian-Med 0.12 released and this will probably be the
> >      version that will go into Etch.  (I do not plan to issue
> >      a further version because we are in the freeze phase - so
> >      dinifti will not be included, sorry.)
This is actually not a problem. dinifti will not be part of etch anyway
(I filed an RC bug against it, because it refuses to work on amd64).

<snip>

> As Andreas said, we did not communicate on our recent work. The release
> will be the prefect moment for advertising our improvements. Maybe we
> could write separately entries for the Debian-Med news, and use them to
> write a nice summary for the Etch release ?
> 
> The above two points are direct TODOs for the Etch+1 debian-med. If you
> like the idea, I propose that we set release goals for the next cycle.
> So for the moment, we would have :
> 
> * Keep in sync the metapackages and the web site.
> 
> * Communicate more regularly to our usres.
> 
> 
> The second point is not really technical, but the first point has been
> discussed a few times in the past. I had a quick look as the sources of
> the metapackage, and it should not be too difficult to write a webpage
> generator from them. I think that it would be a very useful tool, which
> we could expand. For instance, we could also generate
> developper-oriented pages which allow to track upstream versions and
> check the number of installations. Here is an example, made by hand. The
> gool would be to autogenerate such pages.
> 
> http://wiki.debian.org/DebianMedBio (see the popularity and technical
> summary links)
I very much like this idea. IMHO it would be better to let the
debian-med website show what actually is available and not what might be
available sometime in the future. Having this content generated from the
meta-packages itself would be awesome.

The more volatile content might be better placed into the debian wiki as
this removes the bottleneck of a webmaster to a certain degree. 

I know we had this discussion before: A wiki does not increase the
number of contributions per se, but at least for me it would. I find it
much more satisfying to be able to just modify something instead of
posting patches and waiting for someone to apply it.

I'm not saying that Tobias is not doing good work -- because he does
good work. But in fact I cannot really say anything about it, because I 
never submitted a patch (only suggested modifications). This is partly due 
to missing knowledge at my end. ATM I do not even know where I can get the 
wml files from. A wiki is much more intuitive.

Another 'advantage' would be a lower threshold of when some information
is worth putting it on the website. *.debian.org looks and is very
official. A wiki on the other hand is a synonym for work-in-progress. It
therefore has the potential to produce a more dynamic content.

I know that talk is cheap, so I'd volunteer to take care of a
DebianMedImaging page in the debian wiki if there is a consensus that
moving towards a wiki is a good thing at all.


Best,

Michael


-- 
GPG key:  1024D/3144BE0F Michael Hanke
http://apsy.gse.uni-magdeburg.de/hanke
ICQ: 48230050



Reply to: