Debtags and Debram (Was: Re: New debtags suite just uploaded)

Enrico Zini zinie@cs.unibo.it
Tue, 6 Jul 2004 18:11:40 +0100


--Ns7jmDPpOpCD+GE/
Content-Type: text/plain; charset=iso-8859-15
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Sun, Jul 04, 2004 at 08:58:41PM +0200, Benjamin Mesing wrote:

> as a person having no deep insight in the art of categorizing, there are
> some things I do not understand, so I ask if you could widen my horizon
> :-)

Not really your fault: I notice I've been creating and using lots of
special word meanings without caring for people to actually share those
meanings... sorry!  I'll try to take the glossarium out of my head :)


> > I've been browsing through debram.txt, and I saw that these terminal
> > groups really made sense.  A first idea I have for a merge is this:
> I assume with terminal groups you mean the categories in the debram
> system (i.e.
>         1000 User-oriented Packages
>                 1100   General Commands
>                         1110     File Commands
> )?

Almost: I intended the debram categories that are not split anymore.
1653 "Mailbox Handling", but not 1600 "Networking", nor 1650 "Mail".


> >  1) Given the list of all your terminal groups, see if they are all
> >     representable using intersections between different debtags facets.
> >     If not, then refine the facet structure to be able to represent your
> >     categories.
> Hmm... what exactly to you mean by this?=20

For example, take 1653 "Mailbox Handling" and see if it can be tagged
in a way that is both satisfactory and unique among all the other debram
groups.  It may be that this isn't always possible: let's take 1653
"Mailbox Handling", for example:
 - we could use media::mail and use::storing, but we have no applicable
   tag from the mail:: facet, which is not satisfactory
 - if for example there were a "Mail Delivery Agent" group, then both
   media::mail and use::storing could also be used for it, and so the
   tagging wouldn't be unique.

(Note for Thaddeus: I noticed that goldedplus and multimail are
in 1657 "Other Mail", but I'd say they're really mail user agents,=20
fitting better in 1651 "Mail User Agents" instead)


> >  2) After upgrading the facet structure, make a single tag for any of
> >     your terminal groups,
> Creating a tag for each terminal group seems to make sense to me.=20
>=20
> >     then load up the collection in tagcolledit and
> >     try to attach to the groups the proper faceted tags
> Here again I am confused, it seems like you start to use synonyms. What
> exactly are the "groups" and the "faceted tags" here?=20

The idea was to convert the debram groups into something understandable
by debtags tools (like tagcolledit); something understandable like, for
example, a single tag in a special "debram" facet.

After that, tagcolledit could be used to convert from the special debram
tags to the other debtags tags.

However, I now realize that this process could be fully automated: once
we do point 1), we already associated a good set of debtags tags to
every debram terminal group, so it would be easy to create a script that
just adds the tags to all the packages in that terminal group.  The
procedure at point 2) is only useful if we feel like the conversion
should be manually assisted.

> I think it would also be useful if one could find some direct mappings
> of some of the terminal groups to allready existing tags (e.g. 1714 TeX
> to tech::tex) as this would reduce the number of terminal groups to be
> considered.

Yes.  I think in many cases we'll find this to be quite natural, too.


Ciao,

Enrico

--
GPG key: 1024D/797EBFAB 2000-12-05 Enrico Zini <enrico@debian.org>

--Ns7jmDPpOpCD+GE/
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Digital signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFA6t1M9LSwzHl+v6sRAr81AJ9677JwA1EmEDh4ucSmvNTw4hfAqgCeJuFF
Nj7F+T6io9YeiGb8cgDea30=
=KJRI
-----END PGP SIGNATURE-----

--Ns7jmDPpOpCD+GE/--