[Debtags-devel] Re: Recent progress

[Thaddeus H. Black]

--4Ckj6UjgE2iN1+kY
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

Herve's and Erich's conversation has been interesting to
read.  They have covered the issues pretty well, I
think, so I would add only two remarks.

Herve:

> [Erich:]
>
> > As soon as the maintainers will do this, our
> > vocabulary will need to be mostly frozen;
>=20
> I think it's utopian. It will always evolve. Not
> hugely, but it will, and that's normal. Let's deal
> with that instead of requiring a restrictive freeze.

I had used to agree with Erich in this (in fact I have
said so on this very list).  I still feel that Erich
raises a good point, but I hope that he will not mind if
accumulated experience has changed my mind.  The archive
is seen to expand in unexpected and heavily asymmetric
ways.  For example, Java, GNUstep, Ruby, multimedia,
spam filtering, R statistics, Portuguese, KDE and XML
all existed when woody was released, but the numbers of
actual packages in these categories has grown rapidly in
sarge.  On the other hand, Groff, SVGAlib, the serial
port, CVS, GGI, Vi, ham radio, the console, Dpkg and
Octave are mature categories which have not grown much.
Experience sternly and thoroughly contradicts my earlier
bias toward a static vocabulary.  If the vocabulary is
meant to serve the natural hierarchy of the packages,
rather than to impose an unnatural artificial hierarchy
upon them, then one is led to concur with Herve in this.

Erich, do you feel that keeping the vocabulary flexible
poses insurmountable practical difficulties within the
existing system?

Herve:

> But if you ask maintainers to complete tags for the
> packages they manage, most of them won't do it, I
> fear.

Well, experience thus far tends to support Herve's view
in this, but I admit that I am more optimistic.  If we

  (a) present developers with a working, complete,
  practical tag system---a system which the developers
  themselves are using to find the software they
  personally need---and

  (b) tag most of the archive to the point where debtags
  is really practically useful,

then will the developers not want their own packages
tagged correctly?  Will they not press their fellow
developers to tag packages, too?  Will the day not come
when the control file includes a standard "Tags:" line,
by popular demand?

It is hard to get the body of the developers to pay
close attention to tagging packages when debtags data
are still experimental.  However, we have made big
progress toward filling the data in.  Benjamin has
a promising tagging AI up and running.  With Giacomo's
and Maciej's help, I have got the entire archive
debram-tagged; and Enrico has written the autodebtag
script to translate the data into a form we can use.
For sarge+1, I think, debtags will be fully built and
powered, and will be what Debian users everywhere use to
find the software they need.

We have the manpower.  We have the talent.  We have the
motivation.  We have the tenacity.  We have the
patience.  This is going to work.  Our project is good.
Nowhere does any developer raise a voice against it.
They will support it, I think---once we have given them
something practical to support.  We are not quite there
yet.

By the way, who of us will go to Debconf in Finland in
July?  (If I can find airfare within my budget, I mean
to go this year.)

--=20
Thaddeus H. Black
508 Nellie's Cave Road
Blacksburg, Virginia 24060, USA
+1 540 961 0920, t@b-tk.org

--4Ckj6UjgE2iN1+kY
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iEYEARECAAYFAkIjhZUACgkQh3E0gzgBXn5xigCgou5P4ApCrh64fHkFJW2lYTHW
64IAoLYzONBcclftJhK4QdIfHuChqNCF
=WyJF
-----END PGP SIGNATURE-----

--4Ckj6UjgE2iN1+kY--