more thoughts on .htm vs .html

[Ulrik Vieth]

Just a few thoughts before I embark on trying to get some real work done, 
i.e. write up a physics paper that has been delayed for too long already...

1.) I heard that Netscape on DOS/Win might be clever enough to find 
    a file "index.htm" on the local filesystem even if a link points 
    to "index.html".  Whether this means that it does an 8+3 truncation 
    of the requested file name internally or whether it just knows
    about the extension ".html"  as a special case, I don't know.
  
    If the ISOization process would mean that a file "index.html" 
    under RockRidge really showed up as "index.htm" under ISO and 
    not as something completely different, this might at least be
    some sort of kludgy solution to provide some index files in 
    the texmf/doc/ tree.
  
2.) Concerning Joachim's complaint regarding long filenames for TeX 
    packages such as "chapterbib" not being found when stored with
    a truncated name:  I vaguely recall having read somewhere that
    some DOS TeX implementations (e.g. emTeX) might be clever enough
    to try various alternatives, i.e. both an 8+3 or an 5+3+3
    truncated name, when a long file name is requested via \input.  
    GNU Emacs under DOS also seems to get along when the basename
    of an Elisp package is truncated to 8 chars as long as it is
    unique.  However I don't know what would happen to dvips.info
    when the internal links point to dvips.info-{1,2,3}.

In any case, I thing it would be the best approach for a CD targetted
primarily at Unix to use long filenames under RockRidge and simply
rely on the cleverness of DOS implementations to get along somehow
with truncated file names (*if* they are really truncated and not
mapped to something completely different).

Cheers, Ulrik.