Stephanie Losi interviews Charles F. Goldfarb

Interview by Stephanie Losi for Borders Online

How did you get from law school to SGML?

Just lucky, I guess.

Seriously, I practiced law for a while, then joined IBM because I thought some high tech experience would help me attract clients. I was working on a research project for a law office information system. That was in 1969 -- when most information systems were larger than most law offices! We needed a way to let the data be shared among the different applications in the system. The result was IBM's Generalized Markup Language (GML), and eventually SGML.

Back then, what did you think you'd be doing at this point in your life?

I guess I thought I'd be practicing law. I certainly didn't think I'd be writing best-selling books about a technology that I invented that is changing the world.

What do you think the strengths and weaknesses of markup languages are?

The strengths are the intelligence that they add to data, their ease of use, and the fact that they make the data system-independent, portable, and reusable for multiple purposes. The weaknesses are that they resemble programming languages and word processor styles, so it is hard to explain what they really are so that people can get the most benefit from them.

What really excites you about SGML and now HTML and XML?

SGML makes the infrastructure of modern society possible. Our incredibly complex systems and products require massive amounts of documentation -- 4 million pages for a single model of aircraft, for example, which must be updated quarterly. That documentation couldn't be created and managed without SGML. The same is true for the documentation of nuclear plants, oil rigs, government laws and regulations, military systems -- and anything else that is too complex for a single person to understand and that has life-and-death significance. All of that stuff is documented with SGML.

As for HTML, it made the World Wide Web possible, and that, of course, is changing the world in profound ways. I'm proud that many of the same tags that Ed Mosher and I created for the very first GML application are still in use in HTML.

And XML is truly amazing. By allowing businesses to share data easily, entirely new patterns of electronic commerce and business relationships are made possible -- "supply webs" instead of "supply chains", for example.

What's the most rewarding markup project you've ever been involved in?

Working with the late Yuri Rubinsky to create an SGML architecture that helps people with disabilities. It allows Braille, spoken word, and large type versions of documents to be produced transparently as a by-product of normal document production.

How do you think the balance of power will shift between HTML and XML on the Web?

I think they will coexist. HTML is an application of SGML for expressing how Web pages should appear when presented. XML is a subset of SGML with which you can create your own tags that say what the Web pages mean. Those two concepts -- meaning and appearance -- are complementary; both are needed.

In fact, the W3C is currently working on a project to define HTML as an application of XML. It is called "XHTML".

The Web has really splintered from the original unifying HTML standard. Do you think it's possible to return to a standardized Web?

Not at the level of one set of tags for everyone. That is no more realistic than the idea of a single vocabulary for all professions and hobbies. There can be common shared sets of element types and attributes for use in tags, to be sure, and there are groups working to develop those common sets. But they will only be core sets, not the whole universe.

What suggestions would you offer for getting there, and how would consumers benefit from enforced Web standards?

SGML has proven that there can be a unifying standard that allows for differences, paradoxical as that may seem. The secret is that SGML lets you state what the differences are in a standardized way. XML can perform that same function for the Web. Consumers would benefit because of the increased ease of accessing all kinds of information, both human-to-human and program-to-program.

Stephanie Losi is a content editor for Borders Online.

To Charles F. Goldfarb's SGML Source Home Page.

Copyright © 1998 Charles F. Goldfarb. All rights reserved.
From an article at Borders Online