Re: lynx-dev superscript bugs -- Math on the Web

[Steve White]

Yeah, it's been a mess to put math on the Web.

Postscript or PDF is a perfectly reasonable way of presenting
intricate typesetting.  The user might have to get some software,
but it's freely available.  And for academic papers, well, TeX
is good too.

MathML has come a long way.  The latest versions of Mozilla and 
Netscape have it built-in.  There are a few bugs, but by and large,
it works.

But it feels a little elitist to be using MatML at this time.
Furthermore, it's HORRIBLE to hand-code, by an explicit policy
of the MathML team!  (Aww!  Come on guys, you could have done better!)

So, I've experimented with how much math can be put on the Web just
using HTML (and prettifying it with CSS).  This isn't appropriate for
complex academic math papers. But for pages with just a few math 
expressions, or pages that contain only simple expressions, it can 
work very well.

The result is:  A lot can be done, so long as browsers adhere to 
published standards.  

See <http://www.zipcon.net/~swhite/docs/math/math.html> 
( a work in progress ).

The worst problem, in principle, is fractions.  But you can make a 
pretty good fraction so long as tables are supported.  And hey, 
these days, tables are supported even in Lynx!

Correct handling of SGML has been a practical problem.  Another is
support for superscripts and subscripts.

There should be no problem with super and subscripts, since they've
been part of HTML forever, but the support in Lynx is buggy, and
in links it's nonexistent.  We gotta fix that.

It's even possible to make cross-platform sums and integrals.  Once
again, poor support of SGML and super/subscripts puts restrictions
on these.

On Sat, 10 Aug 2002 13:19:52 -0700, Walter Ian Kaye <address@hidden>
>
>When I started working at SLAC, my then-boss was interested in Ping's 
>MINSE which used GIFs yet kept the original formula in the ALT 
>attribute so it wouldn't be lost. My next boss there was enamored of 
>IBM's techexplorer; we had a brief love affair with the browser 
>plug-in (I didn't like IBM's use of <embed> and javascript, so I came 
>up with some <object> code for Lynx compatibility), but the plug-in 
>turned out to be very buggy render-wise (perhaps it just didn't like 
><object>). The general sentiment now is that we should probably just 
>have raw TeX for formulae in document abstracts, and then if someone 
>really wants to render it, they can pass it off to another app or 
>something.
>Sample: <http://www.slac.stanford.edu/cgi-wrap/pubpage?slac-pub-8544>
>
>



; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to address@hidden