On Mon, Oct 24, 2005 at 01:32:08PM +0200, Frank Küster wrote:
[...]
> > This might be an important question in a sensible split.
> > For example the big modularized xlibs was done in parts to
> > help migrating from xfree to xorg.
>
> Please note that we do not plan to "transition" from tetex to texlive;
> as far as we can tell both will continue to (co)exist upstream, and the
> same we want to accomplish for Debian.
This was just as a "more modular design can sometimes help"
heads up.
> But I am not familiar to the
> problems that the X people tried to solve with the modularized xlibs wrt
> this transition;
I haven't followed it too deeply either. But the generic
idea I can see:
Have small packages with clear functionality, that then can
be easily replaced by a "from scratch" new package from the
to-be-transitioned-to package.
> can you elaborate on how this might affect TeX?
It depends on how tetex and texlive want to coexist.
If they're not intended to coexist at all, this does not
affect TeX at all. Each does its own job and decides on how
to rule the world, and conflicts with the other. Job done.
The other extreme: if they should coexist and be
interchangeable and stuff, it might for example make sense
to have bunches of virtual packages, which either of the
both can provide. Like tex-ctan-PACKAGE or tex-xdvi, etc...
So that other packages can depend on the feature they need.
Say tetex-beamer needs pstricks, it would then depend on
tex-ctan-pstricks, which would be provided by
texlive-pstricks and the tetex-*, that is going to have it.
So I think, the question is: What is the intended audience
of tetex vs. texlive?
One part of this answer has already been given: For
build-dependencies, tetex should be the choice.
But what about the rest?
Elrond
