[tex-live] t1code.tex not found
petr at olsak.net
Thu Mar 11 12:03:06 CET 2004
>>I vote to add the clause to TDS: the subdirectory structure of
>>is only recommended, not required. The author of package can decide if
>>the recommended structure is usable for this package or does not.
>This seems entirely the wrong way round. The distribuition
>system should use a consistent naming and directory structure,
>some principle. Authors of packages should not at all be encouraged
>to go off on their own whims!
I am parsing texmf tree maually (using cd, ls, cat, vi etc.) every day
many times for my specific purposes. I have found that the texmf tree
is more and more complicated from one version to another. I understand
that some things have to be more complicated because of global concept
of the TeX distribution and because there are more and more files in
texmf. But I hate the *useless* complications. If csplain files will be
installed into many subdirectories then this si an example of
useless complication. I spend many hours of making my software and
I don't wish to patricipate with my software on useless complication
which I hate myself.
Please, accept the author's wish and revert the files from tex/csplain
and tex/cslatex to its original locations. Please, do it and show a
little respect to the authors of software used in your TeX distribution.
Without authors you would have no software in your distribution.
The files from csplain was located directly under texmf/tex/csplain/
10 years and there is no technical reason to change this location now.
I have a little comments to TDS: IMHO the TDS declares directory
structure principally for better orientation over the files for human
who are using ls, cd, cat etc. These commands are basic UNIX tools for
parsing arbitrary directrory structure. If you say that the
human access is not important but only machine access (by
special software) then I dont understand, why you separe the
information to the files and directories. The whole information in
texmf can be stored in one (gigabyte) binary file and you can programm
the special access-tools (like kpsewhich) to get or upgrade the part
of this information. This approach is very similar to registers in
Windows. I and many other users hate Windows in order to this
On the other hand: you can say that human *can* parse the texmf tree
(by standard UNIX tools for example). In such case the TDS standard have
to be revised. Many useless sub/sub/ directories do the life of human more
difficult but they do not make machine access noticeably easy because
machine is able to get information from one directory with 50000 files
(one beside second) or from one gigabite binary file without problems.
Why and for whom is the TDS dessigned?
More information about the tex-live