A patch to <IMG ALIGN=>

Michael C. Grant (mcgrant@gomez.stanford.edu)
Tue, 24 May 1994 18:49:20 -0700


One of the problems with using in-line images to render equations is that
you can't easily align the baselines of the text with the baselines of the
equations. To see what I mean, look at

http://www-isl.stanford.edu/~mcgrant/equations/usegifs/qcqp2.html

(I don't have transparent gif files yet so the equations have a white
background. This actually helps to illustrate the point better, though).
This is an "illegal" document of sorts, because it uses illegal values
for ALIGN which I talk about below. But, if your reader defaults to
ALIGN=BOTTOM in such cases you will get the desired effect.

For each of the in-text equations, you should notice that they are
shifted up by as much as 5 pixels depending on the length of the
descenders. (I also happen to think that the equations are rendered
poorly. I think that this can be improved considerably on the
LaTeX2HTML end, by using Metafont to generate lower-resolution
fonts, so I'm not so concerned about that yet).

If you read

http://www-isl.stanford.edu/~mcgrant/equations/Welcome.html

and especially the section called "Arbitrary baseline alignment"
for in-line images, you will see my proposed solution to this
problem. There is also a link there to a patch to a SINGLE FILE
in libhtmlw, HTMLformat.c, which adds this functionality. The link is

http://www-isl.stanford.edu/~mcgrant/equations/HTMLformat.diffs

Try the patch out, and look at the example document above once
more. The improvement should be marked! If you want to see a nicer
looking document without the white backgrounds on the gifs, try

http://www-isl.stanford.edu/~mcgrant/equations/usegifs/qcqp2.html

It is the same document, but it uses xbms which are much larger.

I think this is a rather simple enhancement to HTML, as the simplicity
of the patch demonstrates. I really think that in the long run it will
make a big difference for documents with equations, so I hope the
HTML+ discussion group will consider it.

Thanks for listening,
Michael C. Grant
mcgrant@rascals.stanford.edu