Helmut Kopka's interpretation of the TDS

[Ulrik Vieth]

Hi TDS workers,

here comes some stuff for discussion:

The German division of Addison-Wesley has just published the
long-awaited third volume of Helmut Kopka's book series on LaTeX.  
In the first chapter he discusses the organziation of a TeX system
covering among other things the TDS recommendations in some detail.

Although I previously believed that the TDS drafts were reasonably
clear and unambiguous they apparently weren't and left enough room 
for Helmut Kopka to develop his very own personal interpretation
of the TDS, which he now published in his book.

Here are the points in his description that I found questionable 
and worth discussing:

1.) He argues for the re-introduction of texmf/ini (or for emTeX
    texmf/ini/{std,big}) despite the fact that this was taken out 
    in version 0.999 of the TDS draft. (He cites both 0.98 and 0.999, 
    i.e. both release versions that were issued last year.)

    It apparently wasn't clear enough that texmf/ini was taken out 
    in 0.999 precisely because its contents (.fmt and .base files)
    were platform or implementation dependent and therefore not
    shareable unlike almost everything else in the texmf tree.

2.) He seems to suggest that the <mode> level in the fonts tree,
    i.e. in texmf/fonts/<file_fmt>/<mode>/<supplier>/<typeface>
    could be omitted if only a single mode is used at a site.

    I don't think that this was ever intended or suggested, but 
    perhaps this wasn't stated clearly enough.

3.) He makes some suggestions how to organize the sub-structure of 
    texmf/source, which I didn't fully understand.  Whatever it was
    that he was suggesting, it was certainly something different
    from what I'd call common pratice as used on the TeX Live CD.

    This is probably not much of a concern since the sub-structre
    of texmf/source was left mostly unspecified in the TDS draft,
    but maybe it should have been?

4.) He makes some suggestions for texmf/doc which are somewhat
    different from the examples given in the TDS draft and result
    in a more deeply nested structure that more closely resembles
    the organization of the remaining texmf tree to the effect
    of having texmf/doc/tex/{plain,latex,amstex}/<package> rather 
    than simply texmf/doc/latex/<pkg> or texmf/doc/ams/amstex.

    This might be considered as valid alternative scheme subject 
    to further discussion, but it certainly disagrees with the 
    TDS draft and is likely to cause confusion.

5.) He also mentions texbook.tex and mfbook.tex in the context 
    of texmf/doc and states that he isn't sure how to interpret 
    the copyright statement as to whether previewing DVI files 
    of the TeXbook/MFbook would be allowable or not.

    This is probably not a TDS matter but I found it surpising 
    to find such a statement in a book published by a division
    of  Addison-Wesley, the copyright holder of Knuth's books.
 
6.) Finally, Kopka suggest to organize DVI driver utility files
    into a subtree texmf/dviware/{dvips,dvilj,xview,...} rather
    than just having texmf/dvips.

    Again, I don't think this was ever suggested or intended
    and I have no idea where he may have gotten this idea from.
    Besides, I wonder why so many directories would actually 
    be needed since the only drivers requiring auxiliary files
    I know are dvips and the emTeX drivers.

That's about all of his controversial suggestions.  Apart from these
his description of the major TDS areas seems pretty much acceptable.

Opinions anyone?  Any need to issue some sort of clarifying statement?
I mean, a proposed standard such as the TDS isn't worth very much if 
it leaves too much room for book authors or implementors to develop
their very own interpretation which mostly follows the standard but
differs slightly here and there, since that would defeat the purpose
of a standard and we're back to where we started.

Cheers, Ulrik.


P.S. On a completely different subject: Kopka somehow managed to
hyphenate `UN-IX' using german patterns and \left/righthyphenmin=2.
To some extent this reminded me of the TeXbook quotation about 
this program which had hyphenated `God' ...