On May 26, 5:23pm, Ross Moore wrote:
> The point is that if you are using amsmath then you should be using
> the high-powered math-parsing offered by the `math' extension.
> The smaller images that it generates give better down-load times.
> Also, the HTML markup better reflects the logical content
> of the mathematics, much more so than does a single large image.
> Searching on sub-expressions becomes possible.
[Can I take up a (hopefully) polite quibble here?]
I think your goals are noble, but you're way ahead of the curve!
Both the presentation and content should be preserved once MathML is online,
but I can't see the current HTML/image soup being any more logical or
searchable than the image is (that is, not at all!).
The mixture of fontfaces, sizes & weights that you get from mixing <i>, <sup>
and <img> is at best ugly, and at worst confusing (are those "x"'s supposed to
be the same?) And I've seen at least one browser pushed beyond it's engine by
the complicated markup.
I realize the AMS environments, unlike the regular latex ones, really need to
be parsed to be correctly laid out in tables, but would it be possible to still
generate images for ALL the math (as $NO_SIMPLE_MATH normally does) when math
is loaded?
Oh, but the pages DO, indeed, load much faster!
Incidentally, both amstex.perl & amsmath.perl need a fix to accomodate multiple
paths in $LATEX2HTMLSTYLES; At the end of each, replace
print "\nLoading $LATEX2HTMLSTYLES/more_amsmath.perl";
require "$LATEX2HTMLSTYLES/more_amsmath.perl";
with
do_require_package("more_amsmath");
(It's more modular, anyway :>)
--
--
[EMAIL PROTECTED]
http://math.nist.gov/~BMiller/
