Re: More thoughts

On Fri, 28 Oct 1994 wrote:

%%>>How do you do multilingual stuff, or even mathematics, without changing
%%>>fonts?  I guess in an ideal world there would only be one font, with
%%>>all the 300 or so west european letters, perhaps the same number of greek
%%>>and east european letters, then arabic too, maybe 1000 or so maths symbols,
%%>>and then there is chinese...
%%>The HTML+ spec addresses these issues in the "correct" way for a high-level
%%>markup language. It adds the LANG attribute to many tags and adds tags for
%%>math markups. That way the browser implementor can handle these things as
%%>appropriate for the browser platform. This may be as simple as allowing the
%%>end user to specify fonts to use for various languages and for math, and puts
%%>the font control where it belongs, with the end user.

I would be convinced if you showed me how to do a blackboard R---it looks
like IR with its right edge doubled too---in HTML+.  Also, several
hundred other special symbols that exist in the American Mathematical
Society fonts, with names like gtreqqless, curlyeqprec, ngeqslant,
ntrianglelefteq, varsupsetneqq, ...

%%>>Anyway, you just use that font and slant it, embolden it, magnify and
%%>>shrink it, and that's all there is to it.  Simple eh?
%%>It is if your browser can support these things. Browsers for simple text
%%>terminals (e.g. VT100, 3278) can't.

I was being facetious; VT100 wouldn't even produce the symbols.

%%>>I think we do need font control somehow, and that is why TeX is so
%%>>complex.  And existing fonts so infuriating!!
%%>I think we should avoid font control at all costs.

This still begs the question. To support maybe 40 simultaneous fonts on
screen, the browser is going to have to be pretty complex.  (La)TeX is
the de-facto standard for setting maths on screen and printed, and you
are lucky enough to have Leslie Lamport on your list!  Listen to what
he says about these issues, as he knows far more than anyone else on this
list about such things.  For instance, we have learnt the hard way that

--  font attribute changes like emboldening should work independently in
    text and maths mode
--  no matter how many markup constructs you build in, there is at least
    one that someone will want to use that is left out
--- some markup should be independent of the user-specified font family,
    and some should depend on it
--- the standard postscript fonts are inadequate for maths, and when the
    symbol set is large, WYSIWYG is more of a hindrance than a help, as
    even selecting just one symbol is a pain

Mike Piff

%% Dr M J Piff, School of Mathematics and Statistics, University of %%
%% Sheffield, UK. +44 114 282 4431   e-mail: %%

Received on Friday, 28 October 1994 15:07:38 UTC