[tanmoy@gita.lanl.gov: CTAN archives]

[Sebastian Rahtz]

David Carlisle writes:
 > previous versions if the current one is not backward compatible. Even
 > after thinking long and hard, I could not find any way to automate the
 > process of upgrading our LaTeX distribution. The base files are fine:
 > but the contributed software seems to be a mess.
well, the key word is "contributed"; thats why there is a distinction
between what the core team  write and maintain, and what others offer
out of goodwill. 

 > 1) Could one set up a mailing list where _all_ upgrades to the
 >    packages (base/packages/contrib-supported/contrib-others) will
 >    receive notification. This will at least allow me to upgrade only
we have ctan-ann for this purpose, though its not used enough

 > 2) A `quote site list <date> <root>' command at the CTAN
 >    sites which lists for me every package in the directory tree under 
 >    <root> that has changed since <date>?
we can do this, i guess. if someone else wrote the script, it would
get done faster! (work on the basis of processing FILES.byname)

 > 3) Could the packages be made more consistent? i.e. Require that every
 >    supported package have a .ins file with the following properties:
we're still awaiting adoption of the guidelines, i am afraid

 >               file.cls  LATEXINPUTS
 >               file.sty  LATEX209INPUTS
 >               file.tex  LATEXINPUTS
 >               file.fd   LATEXINPUTS
 >               file.mf   LAMFINPUTS
 >               file.bst  INDEXINPUTS
 > 
 >         It turns out that relying on the extension is not always the
 >         best.
 >      b) Instead of (a) a Makefile which includes ../../paths.mk (or
 >         something similar) and uses no hardwired program or path names
 >         (i.e. uses $(LATEXINPUTS), $(LATEXPROGRAM) etc.) is an
all thats needed is a defintion of TEXMF, and then follow the TDS
guidelines

 > Could you please pass this on to the right people?
hang on, there *are* no "right people"; the LaTeX community is all of
us, not a  committee. the CTAN maintainers could publish guidelines
and refuse things that didnt meet them, but we have no resources for
testing submissions.

you might be amused, tanmoy, to hear that i am creating a Unix
readytorun CD for TUG; among other things i wrote a Perl script to
process all latex contrib packages consistently and install them in a
TDS layout. it can be done, even with what we have now.

if we want a well-managed LaTeX world, we have to invest time and
money into managing it. This group (the TDS) represents many many
hours of people's free time,  and i suspect that in many cases it has
used up their mental allocation of time they can spend on such
matters. the TDS was a vital stage in the development of a rational
LaTeX world, now we have to start over again and develop guidelines
for authors. Can i be the only one who simply feels *tired*? this is
too big an operation to do in people's spare time.

tanmoy could always ask paul ginsparg to put some of his NSF Grant
towards maintaining the LaTeX stup, of course :-}

sebastian