Re: Thoughts on scientific HTML documents

Dave Raggett (
Tue, 31 May 94 12:07:35 BST


Many thanks for expounding on your thoughts the needs for scientific
HTML documents:

I will give my reactions below on a point by point basis.

> A little while ago, I posted an article on Usenet with a proposal for
> adding a "postscript package" extension to HTML readers that would
> serve to improve the state of equation rendering in HTML documents.

The current view is that browsers should be designed as a collection
of "applets" rather than a monolithic whole. We are interested in
allowing these to present different types on info within subwindows
in the document. The FIG element can then be relaxed to support a
wider range of info types, including encapsulated postscript. The
main HTML application and the postscript viewer negotiate on the size
of the subwindow, e.g. to make it fit within the current document
margin settings. The next step is to pin down a portable API for
how these applets communicate with the main app.

Phil Hallam Baker of CERN has a working implementation of HTML+ math
which he has tested on a wide range of math notations. Phil's work
will be fed back into the next revisions to the spec. As a result
of the WWW'94 conference, HTML is being standardised as HTML 2.0 for
the de facto standard given by Mosaic, and HTML 3.0 for the features
developed under the HTML+ banner. The math support will be included
in the 3.1 release to give us time to review Phil's work. I hope to
include support for math in my reference browser to be released this

With regard to your needs for better control of the vertical alignment
of IMG elements, I have already added a BASELINE attribute to IMG for
this purpose. Support for this will be included in browsers that comply
with HTML 3.0.

I am working on a proposal for scaling attributes for IMG and FIG which
will allow browsers to scale inline images, based on either the current
margin settings or as you suggest the font size. The HTML 3.1 math
element will allow equations to be scaled by scaling the fonts.

On how to handle images on text-only displays ...

The FIG element is designed to cater for graphical and non-graphical
displays in a clean consistent way. We may also want to add a new
mechanism for offering better support to the visually impaired, e.g.
by an attribute giving the URL of a spoken explanation. Similar
support is needed for the MATH element.

On reducing connection overheads for inline images ...

There have been proposals for sometime for using multipart MIME messages
to send the images with the document text. Caching complicates this as
the client may have cached some but not all of the images. This can
be handled as a multipart GET. The HTTP spec needs to clarify these
techniques and browser writers encouraged to support them.

Best wishes,

Dave Raggett

----------------------------------------------------------------------------- Hewlett Packard Laboratories email: Filton Road tel: +44 272 228046 Stoke Gifford fax: +44 272 228003 Bristol BS12 6QZ United Kingdom