Standardization Effort Surrounds Web Language

HTML specification to be published by IETF

An ongoing effort to standardize the World-Wide Web's text-tagging language is now centered on the Internet Engineering Task Force (IETF). The standardization effort concerns the HyperText Markup Language (HTML), the standard set of tags used to prepare text for use on the Web, which is used by anyone preparing documents to be placed in a Web server and accessed by Web clients.

HTML is one of a set of languages comprised by Standard Generalized Markup Language (SGML), the data encoding system that allows information in documents to be shared-either by other document publishing systems or by applications performing electronic delivery and other functions. SGML is a vendor-neutral, formal international standard, as HTML is intended to be.

Tim Berners-Lee originated HTML as he was developing the Web. Since then, a leader in standardization efforts for HTML has been Dan Connolly, software engineer for Hal Software Systems, Austin, TX. As the Web became popular, Connolly discovered HTML on USENET, the worldwide UNIX bulletin board, where Berners-Lee had posted it. "I tried to come up with a document type definition-an SGML specification, basically-and it fit like a square peg in a round hole," Connolly says. Eventually, others wanted to know what was legal and what was not using HTML, so Connolly continued his standardization effort, trying to retrofit HTML to SGML. A draft HTML specification was published for the first time on the Internet in June 1993.

As items such as footnotes, superscript, and text centering needed to be added to the basic tags for headers, lists, paragraphs and links, a new specification called HTML Plus came into use, beginning in Bristol, England. Since March of this year, Connolly and Web and Mosaic developers, have been trying to produce a published specification through a working group of the IETF. "We kept spinning iterations and trying to get it just right," Connolly says. The specification is scheduled to come out as an Internet request for comment (RFC) from the IETF late this year. Once the specification is published, the IETF will be responsible for approving new features. "We're doing the best we can just to get this standard, that describes what's been going on for months, out the door," Connolly says.

Developing the standard has been a job of delicately balancing actual coding practices with the SGML standard itself. Both are important. "What people are doing is important so that when somebody goes to implement a new browser, he'll be able to use all the data that's out there," Connolly says. "The other side is to try to use SGML because it's an ISO [International Standards Organization] standard and there are benefits to be gained. For example, since we used SGML, the documents can use readers for blind people where you can take an SGML document and feed it in and it will read to them."