[Debichem-devel] Chemistry wiki page

[Andreas Tille]

On Tue, 7 Oct 2008, Daniel Leidert wrote:

>>    1. Create tasks files and cdd-dev stuff under
>>       svn://svn.debian.org/cdd/projects/debichem/trunk/debichem/
>>       (or whatever name you prefer)
>>       --> I volunteer to do this provided somebody takes over the
>>           categorisation work
>>    2. Make science-chemistry package simply dependant from meta
>>       packages debichem-* (as it is done in science-biology).
>>       --> To enhance the tasks pages I have to work down my todo
>>           list for the task page creation code to resolve these
>>           dependencies on the web page (it is quite high on the
>>           TODO list)
>
> Ok till here as long as someone cares about it.

I just offered to care about - as long as some chemist takes over
the categorisation.

>>    3. Enhancing DebiChem team by people who volunteer to work
>>       on categorisation stuff and hunting for prospective packages
>
> But I'm not ok with this one and I sent Andreas a mail in private. We
> don't need people to search for prospective packages. This will just put
> more and more work on the shoulders of those few people, who maintain
> the packages.

"Enhancing DebiChem team by people" will put more and more work on
those few people?  Could you please elaborate on this?

> This is not an enhancement to the team and never will be.
> If people are willing to care about the packages, then they can add
> them. I don't want to get into the same trouble, Ubuntu put us by
> uploading any and every package without making sure, that someone will
> care about it. I adopted several of those once-uploaded-never-maintained
> packages.

Well, I did the same for the Biology packages.  The main difference
is that I was able to increase the team of people actively working
on this and thus I was able to lift burden from my own shoulders.
This is the message I would like to send - but the price you have to
pay is to invest some time into forming a real team.  If you don't
tell the world that you are caring for chemical software in Debian
and just let them stumble about it by chance you can not expect that
many people will join.  When starting the Debian Med project in 2002
there was nearly nothing except some biological packages hanging
around.  If you look at the graph

    http://people.debian.org/~tille/talks/200808_med/authorstat_med.pdf

which shows the main activists of the Debian Med mailing list you can
see that it took about 4 years until the term "team" is rectified.
So it just neads patience and effort to form a team.  But my intend
is to spread the idea and experience we have won there and so your
situation could be much better.  The big chance is that we now have
tools you can easily impress people and simply show what you are
working on - this is a very convincing way to attract people.

>>    4. Advertising DebiChem as Debian internal solution to make
>>       Debian the distribution of choice for chemistrists
>
> If there is man-power to do this, then go on, but in general do *not*
> expect, that the package maintainers will take over this task too.

Well, you might decide whether you are actively seeking for help or
just continue working.  In the Debian Med case this search was successful
to share the work (and as you might have noticed in the Graph the
cooperation between projects might be considered successful considering
the fact that you are number 8 of most active posters).

> I told you, what happened to chemistry related distributions.

There is no need to tell me - it is the same with biology related
distributions.  That's why we do not build a separate distribution
but stay inside Debian.

> 2 team members and 26 packages (+ 1 or 2 ITPs) and both team members are
> also part of several other groups with several other packages or ports.

That's why my intention is to add more members, but you obviousely
did not realised this intend.  It is not only about managing single
packages - it is about managing the maintainance of a set of packages
sharing work between people who do not even realise that they are
potential coworkers.  It is sometimes about involving interested people
for instance upstream.  Compare two situations where you write an
e-mail to upstream:

   1. I'm Daniel Leidert and would like to ask you whether you like
      to do some work to build your nice software project as Debian
      package.

   2. My name is Daniel Leidert and I'm writing you on behalf of the
      DebiChem team which tries to build a complete framework for
      scientists working in the field of chemistry.  We have reached
      the following state <Link to information about your project>
      and our agenda says <Link to your future plans>.  If you are
      interested top propagate your software project inside this
      framework we would like to support you with any technical help
      to build packages, sponsor your packages etc.

Which way would you consider more promising to get a Debian package
with less effort for yourself?  There is no guarantee that 2. will
work - but it has worked in the past for us and we were able to
involve people who do a certain amount of grunt work packaging which
leaves us the tricky bits for polishing and happy people on the
other side.

> <rant>
> What should we do? Concentrate on task pages and let the bug count grow?

Nice try.  You noticed that the tasks pages were only the first step and
that I'm working on QA related stuff?  QA related stuff which might be
helpful for projects with the same intend to create a complete framework
and are not focussed on a single package view.

> Are you serious?

Yes, perfectly.

> What about if you sent me some skilled maintainers for
> the Debian XML/SGML packages?

Off topic, and yes, I have also other packages I would need some
help with.

> Maybe then I can concentrate a bit more on tasks like renaming "CDD"?

Even nicer try. ;-)
Renaming CDD is the most boring and stupid task I was ever undergoing
inside Debian.  If you would have hold about 10+x talks about CDD stuff
and you was asked why you are creating just another distribution even
if you are told that you are doing not 15 minutes before explicitely
you would also undergo this task before going mad.  So either I will
finish the renaming stuff or end up in the hospital (and are not available
for any packaging any more at all ;-)).

> </rant>

So this was a quite weak rant.  Please try a better one.

> Regards, Daniel

... and BTW, I always liked your technical help regarding automake
stuff etc. and so I really reward your technical experience.  I tried
to give back with helping in managing a subproject.

Kind regards

        Andreas.

-- 
http://fam-tille.de