Debtags for defining the minimal age that a program can generally be used

[Miriam Ruiz]

2013/9/12 Andreas Tille <andreas at an3as.eu>

> Hi Miriam,
>

Hi! :)

On Wed, Sep 11, 2013 at 11:20:44PM +0200, Miriam Ruiz wrote:
> >
> > Maybe we should start setting up the lists of packages that might be
> > relevant for kids, at the same time.
>
> Well, the list is setup here
>
>    http://blends.alioth.debian.org/junior/tasks/packagelist  [1]
>
> and it is also categorised here
>
>    http://blends.alioth.debian.org/junior/tasks/             [2]
>
> > It's not obvious what those packages
> > might be, I mean, there's like a dozen of them which are very clear,
>
> Could somebody please check those who are clear and if they are missing
> please give me a ping in what task it should be (or what task might be
> wrong).  I'd happily volunteer to do the according
>
>   echo "Depends: <pkgname>" >> <taskname> ; svn commit -m "add <pkg> to
> <task>"
>
> if somebody would simply tell me.  If you prefer Git please tell me
> right now and you will have the Debian Jr tasks in Git in 24h.
>

SVN is okay, thanks :)

I can do that, and it would be a very visible first step, but in my opinion
that would be like starting to build the house from the roof. In my
personal vision -that of course is open to debate, if anyone is willing to
share theirs-, the task of the Kids Team is essentially a classification
one. Of course, more tasks can be added, like packaging new things,
spreading the word, polishing existing packages, and stuff. But what I
would like to have is some metadata regarding at least a relevant subset of
the packages in Debian, including relevant information about whether those
packages are suitable for which kind of kids. That's my main goal.

As a side note, I would like to open a debate about this, if anyone sees it
differently and wants to share their point of view.

So, to achieve that goal, I would like to have somewhere to store the
metadata needed for this classification. DebTags has always been the most
obvious possibility. Another option discusses was to add an extension to
desktop files. The most quick and dirty option would be to set up an sql
database somewhere and export everything from that. If DebTags are out of
the equation, any suggestions about this?



> > but more in depth analysis is going to probably be needed for the rest.
>
> +1
>
> > I'm not
> > really sure about how to handle that, without relaying on some
> > community-based tool such as debtags or wiki, and I obviously don't feel
> > myself capable of doing it all alone.
>
> You are not alone on the technical side but I'm simply lacking in
> knowledge.  I do not think that debtags should be used to design a task
> in terms of a set of packages and to my surprise Enrico told me exactly
> this at DebConf.  So if you don't believe me - please ask "Mr. DebTags"
> for what purpose he invented this stuff.
>
> If you prefer a Wiki over Blends tools - feel free to even use a Wiki
> and I might watch it to port the diff to the Blends tasks files.  However,
> I wonder why you
>
>   a) want to duplicate the work that is just done (see [1],[2])
>   b) think that the one liner above is harder to do than a Wiki edit
>      for you / my offer to work as "one liner proxy" for people
>      not comfortable with SVN / Git
>
> > > You should expect some delay / send some ping to the debtags list
> > > because it is not very active since some years[1].
> > >
> >
> > Yup, I guess that we will have to be patient about this.
>
> So, why not starting right now with something that does not need this
> level of patience?  I'm honestly trying to find out what people keeps
> away from working with tasks files.  I'm sure all the people who are
> *currently* involved are not afraid about simple d/control like text
> files and working in SVN or Git.  So this can't be the explanation.  For
> others I might consider a GSoC 2014:  Webdesigner for Blends tasks.
>

We certainly could come up with a quick and dirty list of suitable packages
to set up the tasks, I don't think that should be a problem. The main
problem is that instead of coming from some debate, consensus or even
community knowledge, I would just come out from a list of packages that I
would think appropriate for each age rank. I certainly think we can do
things better, at least if other people want to be involved in this. The
good point about starting with the tasks is that we would already have
quick results that might motivate other people. In my plan, there were some
initial stages before defining some subsets of packages for each tasks.
Like defining the tasks themselves, for example.

Greetings and lost of thanks,
Miry

PS: I'm keeping debtags mailing list CC'ed in this mail, because they might
be still be interested in some part of this mail, but I'll probably remove
them from CC in the next ones
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.alioth.debian.org/pipermail/debtags-devel/attachments/20130912/30c400a7/attachment.html>