INFORMATION TECHNOLOGY

Group That Sets Web's Technical Standards
Offers System for Displaying Math Symbols

The World-Wide Web has never been very good at math. Mathematical symbols and equations have been difficult to display on line because standard text formats made no provisions for rendering them.

But on Tuesday, after more than two years of work, the group that sets technical standards for the Web recommended a set of specifications that will let Web pages display everything from pi to the quadratic equation (below).

The solution is MathML, or Mathematical Markup Language, a series of codes that can be used to display mathematical symbols and equations on Web pages. It was developed by the World Wide Web Consortium, an international group made up of more than one hundred universities, corporations, and research organizations. MathML supplements HTML, or Hypertext Markup Language, the standard coding used to create Web pages.

People who wish to use the new markup language can download a "plug-in" that will interpret the codes for their Web browsers. (A list of such plug-ins is available on the consortium's Web site.) A plug-in is an ancillary piece of software that allows a Web browser to perform an additional task, such as play music or show a motion picture. Many software makers and journal publishers, including a number that helped develop the guidelines, have pledged to support the new standards.

Until now, many scientists and mathematicians have usually had to convert symbols and equations to image files in order to include them on Web pages. Besides being cumbersome and time-consuming to create, image files make it difficult for readers to search for mathematical symbols or specific equations, because browsers cannot easily tell what kind of information an image file contains. MathML will allow the encoding of mathematics symbols, equations, and formulas in a way that browsers can search, manipulate, and display.

"It's something completely new," said Vincent Quint, a research director at the French National Institute for Research in Computer Science and Control who is a deputy director of the consortium. "It's not only about displaying equations and printing equations," he said, adding that MathML will also allow scholars to communicate with one another and to improve their teaching.

Robert R. Miner, co-chairman of the consortium's Math Working Group, said that MathML would let a scholar search a data base for a specific mathematical equation, move the equation into a math-editing program to try to improve it, then paste the equation into an e-mail message to send to a colleague. "That kind of seamless cut-and-paste from one application to another is extremely difficult now," he said. "But that kind of free flow of information, I think, will have a huge impact on both research and teaching."

MathML could be a boon to on-line scientific publishing as well. "One of the major problems for scientific publishing, especially for mathematics, is that HTML does not supply mathematical typesetting capabilities," said Andrew M. Odlyzko, head of the department of mathematics and cryptography research at AT&T Labs. Besides creating image files, he said, on-line scientific journals have had to use "very clumsy procedures" that required readers to load special software and thereby limited the audience.

The researchers who developed MathML hope the system will help blind Internet users as well. The markup language could be used by a speech synthesizer to read mathematical equations aloud. "I think the concept is very good," said Jon Gunderson, coordinator of assistive communication and information technology at the University of Illinois at Urbana-Champaign. He said the system may be slow to catch on, however, because most Web browsers cannot decode it without a plug-in. "My concern is that people for the short term will continue to use images to present equations," he said, because browsers lacking a plug-in will still be able to display such images.

MathML is compatible with Extensible Markup Language, or XML, a new coding language for Web pages that uses a more-complex way to code and structure data.