compressed doc files and apache

[Douglas Bates <bates@stat.wisc.edu>]

I very much like the new tetex-doc package that puts so much TeX
documentation under /usr/doc/texmf.  I run apache on one of my Debian 2.1
systems and was going to tell my students to start looking for TeX
documentation under
		http://machine.name.domain/doc/texmf/

That works fine except that many internal links are broken.  The link
to doc/texmf/etex/base/etex_man.dvi, for example, doesn't come up with
the dvi file because the file has been compressed with gzip and is
stored as /usr/doc/texmf/etex/base/etex_man.dvi.gz

Is there a way to configure apache to look for the gzip'd version of a
file if the original file is not available?

Also, would it be reasonable to allow .pdf files in the /usr/doc tree
to be uncompressed?  That is, put the original .pdf file in the
package and not a .pdf.gz file.  My understanding -
correct me if I am wrong - is that a .pdf file can have different
levels of compression internally.  I think there are settings for
pdflatex that change the amount of compression in the .pdf file, for
example.  If we required a high internal compression level in the .pdf
file I imagine that then running the file through gzip would not
produce substantial additional savings in file space.

The gv viewer is able to handle a .pdf.gz file but acroread is not,
AFAIK.  This means that a .pdf.gz file has to be copied to a temporary
location then gunzip'd if you are going to view it with acroread.