[tex-live] Perl for Windows

[gnwiii at gmail.com]

On 6/21/06, Hans Hagen <pragma at wxs.nl> wrote:
> [...]
>
> ==== intermezzo
> [...]
> ====== tex
>
> so, how about tex (and the future) then
>
> i got the feeling that which each year of tex live the 'let's make something
> great for all platforms and treat them equal' were coming more close to
> 'tex live is for linxu geeks only' (the split in win/lin code base is the most
> prominent sign of this)

In this sort of project, you get what people are willing to provide.
I'm not surprised that fewer people are willing to write code for
Win32 than for linux.

> i wonder if it wouldn't be better to think about:
>
> - texlive as a tds complient resource thing only
> - keep the binaries (and platform things) out of the game
>
> that way we can have
>
> - tex live + linuxtex (vendor dependent variants)
> - tex live + miktex
> - tex live + ptex (pretty good cjk distribution)
> - tex live + cross-platform-tex
>
> etc. it saves a lot of discussion, we no longer need to maintain the
> same binary set for each platform, it would open roads for (indepedent)
> binary installers, etc
>
> it's not so different from what happened last year
>
> (to be honnest, i think that the tex community is not (or no longer)
> capable of thinking cross platform)

If all the above can really work with a common texmf tree, this plan
makes lots of sense.  In practice the installer needs to tell the
user:

"found pdfetex version 1.20 in the path! Please update to version 1.30
or higher or adjust your configuration"

It would be good to maintain a list the TeX binary packages with which
the TL tree is expected to work.  This would give certain linux
vendors an incentive to update their tetex packages and provide a URL
other than "how to report a problem" in replies to certain bug
reports.

What we loose:

1) the ability to promote the spread of useful new binaries (e.g.,
sam2p, xetex).
2) control of versions of texmf/scripts that were copied over to
    the binary distro in years past (rather than use the kpsewhich
    method to find scripts in the texmf tree at runtime)
3) uncertain support for legacy unix systems (now that tetex is not
being maintained)
4) opportunities to complain about certain OS's

There could be some benefits to people trying to put together new
linuix distros -- rather than deciding what to include on the basis of
user requests, the distro can
provide a really minimal texmf tree (e.g., using system fonts) and
suggest to people who want more that they install the TL tree.

-- 
George N. White III <aa056 at chebucto.ns.ca>
Head of St. Margarets Bay, Nova Scotia