Helmut Kopka's interpretation of the TDS

Paul A Vojta twg-tds@mail.tug.org
Mon, 25 Nov 1996 23:06:47 -0500


> Date: Mon, 25 Nov 1996 11:58:25 -0800
> From: mackay@cs.washington.edu (Pierre MacKay)
> To: twg-tds@tug.cs.umb.edu
> Subject: Re: Helmut Kopka's interpretation of the TDS
> 
>    o    TeX and its support programs become bigger and slower.
>    o    If you want to look directly at a macro file (e.g., plain.tex or
>         amstex.tex) (which I do from time to time), then you have to dig
>         to find it.
> 
> This is giving a name a bad dog.
> 
> The TDS specifications do not do anything to the size of TeX, they
> do seem to create a few more inodes on a Unix system than the more
> rational ;-} way of setting up the tree, but they say nothing about
> the size of TeX itself.  

I beg to differ.  Quoting:

TDS> As a result, the \abbr{TWG} concluded that a comprehensive \abbr{TDS}
TDS> required that implementations of \TeX{} must support some form of
TDS> implicit subdirectory searching.

> The inclusion of the kpathsea library does add a bit:
> I show -rwxr-xr-x   2 texmf    adm       101284 Oct  5  1995 libkpse.so
> for the shared object library, and 147000 bytes for the static library

That's what I was referring to.

> With an up-to-date ls-R database, I don't sense any serious
> delay introduced by kpathsea at all.  I very much doubt that
> a huge TEXINPUTS etc set of environments such as I used to use
> would make things any nippier.

No, but a short TEXINPUTS resulting from putting all the input files in
one directory would (at least from a user's point of view).  See my
previous e-mail.

--Paul Vojta, vojta@math.berkeley.edu