[tex-live] t1code.tex not found

[Fabrice Popineau]

>> You can disagree but that's what the TDS specifies :

> I *know* that this is the consequence of some stupid decisions of
> authors of TDS, because they put out of mind: "the power is in the
> simplicity".

See below about simplicity. Depends about 'simple for who' ;-)

> I don't plan any sharing of files from tex/csplain with another
> external programs (the example about pictures given by you is
> irrelevant).  I vote to add the clause to TDS: the subdirectory
> structure of tex/package is only recommended, not required. The author
> of package can decide if the recommended structure is usable for this
> package or does not.

I would vote against it. TDS is not tight enough, I certainly don't want
to see it more lousy.

I want to see a TDS where you can locate each places belonging to some
'foo' package, and that all those places do not clash with any 'bar'
package.

> The users can be confused if the etalon csplain.tar.gz has another
> directory structure than your TeX distribution. I wish to remove this
> users confusions. This is main reason of my movements. 

I can understand that. But imho this should stay hidden to people. And
I'm sure the vast majority of people won't dig into this directory
structure. The whole business about this is to reach a point where we
can automate the distribution of ready-to-run packages. At this point
they really won't care anymore because it will be easy to remove the
'foo' package and replace it with a newer version without looking into
the files.

> Is there a technical reason why csplain have to have the "base" subdirectory?

Yes, it makes maintenance over the whole texmf tree _simpler_.

Before I moved a couple of files like these ones, when you wanted to get
all the files for some texmf/foo/bar/* stuff, you couldn't know if you
had to recursively search into the /* part or not. It depended of the
package. Now I have a small set of rules that comply with my requirement
above. But they do need to apply the TDS more strictly.

The current status of the texmf tree  is unstable: I'm  working on it. I
intend to sort it  out as far  as I can and  certainly before the end of
the week. The texmf-dist/  part is almost  ok to me. This is a  _proposal_ for
something easier to  maintain  than what  we had  up to now.  However, I
agree that    the majority of people  here   may reject  it  for various
reasons. Without a better scheme, that would be a pity in my opinion.

Greetings,

Fabrice