However, I think there is at least one *serious* reason to consider a
client-side extension standard: namely TeX. For mathematics,
physics, and other technical disciplines, HTML is just fundamentally
inadequate (with all due respect to the impressive LaTeX2HTML
converter). People in these disciplines have just as much need for
hypertext as everyone else---for example, references to other online
papers, or inline images/animations. What they need is "hyper-TeX".
Now, it seems unlikely that DVI previewing will be built into Mosaic
any time soon. Aside from the attendant code bloat, DVI previewing
requires a lot of external resources (i.e. fonts), which would badly
break the self-contained model that Mosaic wants to present.
A more realistic scenario is for NCSA and the Web community to define
a standard interface (analogous to the CGI standard for server
extension scripts) which allows client-side external viewers to act
as Hypertext Extensions.
First, such a standard should specify a way for external viewers to
send URLs back to the WWW browser that spawned them. Second, this
standard should allow the viewer to specify whether it wants to
handle the result of accessing the URL itself, or let the browser
handle it (to allow external viewers to perform their own inlining).
With such a Hypertext Extension standard in place, it would be
relatively easy to implement hyper-TeX, since the DVI format already
supports a system of external references through its \special escape
mechanism (currently used to inline local images). The \special
format could be extended to encompass hypertext anchors, which would
be displayed visually by the DVI previewer.
I recognize the major dangers of opening up the client side---in all
likelihood we could just end up with the system incompatibility
nightmare from which the Web has so far let us mercifully escape.
But if TeX is going to be accommodated (which I think is necessary
and inevitable), then we should consciously decide whether it's going
to be a one-time hack or if we want to take the opportunity to define
a general extension mechanism.
--------------------------------------------------------------------
Paul Burchard <burchard@geom.umn.edu>
``I'm still learning how to count backwards from infinity...''
--------------------------------------------------------------------