Joachim Schrod TWG-TDS@SHSU.edu
Fri, 22 Sep 1995 13:50:54 +0200 (MESZ)

Karl wrote:
>     > The question, as far as I know, is whether the TDS should require
>     > implementors to provide a way to simultaneously define one path for one
>     > program (latex-foo) and another path for another program (latex-bar).
>     Yes, we should. 
> Joachim, I'm afraid don't understand. What is the proposed change here?
> Is the idea that we're just clarifying that this is the case? That
> filenames need not be unique across different formats? There's no
> ``real'' change to the directory structure going on here, just better
> explanations of the ramifications?

Yes, that's my point. Describe the fact leading to a requirement, not
a technical solution.


As I've said in previous mails, the TDS document should not describe
implementation strategies, but should just describe layouts and add
the implications for requirements if we detect from our feedback that
the implications are not clear for our readers. To get the proposal
accepted by the TeX community, we have to make sure that the
requirements are implementable with the systems today. If the
implementation is not obvious, we should record that check in an
appendix, for our reader's information.


There is an obvious implementation of the requirement for different
TeX input files search paths, namely wrapper scripts that set that
path. Most TeX implementations use some script (or configuration
file, in the case of GUI-based systems) anyhow to call TeX, since
they need to record the format somewhere.


Joachim Schrod			Email: schrod@iti.informatik.th-darmstadt.de
Computer Science Department
Technical University of Darmstadt, Germany