On W3C, The Web, Information
The great dragon Obrfynar climbed the side of the mountain; reaching the podium, she turned to face the audience. There was polite applause, curtailed because everyone wanted to listen. She was, after all, one of the most famous dragons on the conference circuit. Mountain birds flew hastily away in case she cleared her throat, but she just took a drink of sheep’s blood from the glass on the podium and began to speak.
“We are Dragon-kind! All look to us in awe! We are the mightiest! And from the beginning of time we have known what we want, what we need, what we must have, what we will have! And what is it?”
There was a predictable roar from the crowd. Any birds foolish enough to be watching had long fled or were roasted in the excitement. The roar said one word:
“We must seek gold. Wherever it is, we must find it and hoard it!”
I first heard about the World Wide Web in 1991. I’d released a text retrieval system as open source (lq-text, originally a commercial product, nx-text) and someone from the University of Toronto contacted me about using the “WWWW format” for archives. If I’d realized the Web was using SGML I’d have been very excited, since I was working at SoftQuad in Toronto, and we made one of the best-known SGML editors.
It wasn’t until 1993, though, when I encountered NCSA Mosaic being demonstrated by its two young authors, Marc and Eric, that I realized the importance of this new invention. The demo was at the ACM SigIR conference, and the organizers interrupted the conference schedule to include the demo, knowing how important it would be. They were right.
I came back to Toronto and we got Mosaic running and showed it to Yuri Rubinsky, the charismatic president of SoftQuad.
Yuri and I flew down to NCSA to meet with Joseph Hardin at NCSA. Joseph wanted us to make an editor for Web files, for HTML. Why? Because he felt it would legitimize the format as being derived from SGML, an international standard. Those two young men who wrote Mosaic and who were now starting their own company, Joseph told us, wanted to move the Web to using Rich Text Format, based on a standard published by an industry consortium rather than ISO, and, more importantly, losing the accessibility and the separation of form and content.
While we were visiting NCSA, Yuri also took time to argue for table headings, so that wide tables could be automatically transposed for publishing in Braille. There were people opposed to marking up table headers: a reminder that accessibility has always been a fight.
“But it is not only gold that we need,” Obrfynar continued. “It is glittering gold. And gems and sparkly things. Wristwatches and buckles and pacemakers and splints! Not dirty silver that tarnishes and grows dull but shiny titanium steel and glowing uranium and gems and above all else more gold!”
An early competitor to the Web was HyperG from Graz university. In order to make your server visible it seemed that you had to contact the administration to be told where your server would fit in their subject hierarchy. The Web succeeded in part because anyone with an Internet-facing server could put up a Web site; if you had administration experience it took maybe an hour or two to get something going.
But at the same time that the Web was growing, SGML usage itself was growing: encyclopædias, dictionaries, major reference works, newspaper metadata, large technical manuals (some with over 100,000 printed pages) were either in SGML or moving to it, away from proprietary formats. And now people wanted to be able to put those SGML documents on the Web.
It happened that a colleague brought back a paper from a conference that got my attention, and before long SoftQuad was shipping not only HoTMetaL, the HTML editor whose development I’d managed, but also HoTMetaL Pro, and now SoftQuad Panorama, a Netscape plugin to view SGML files on the Web.
Unfortunately, another company also had an SGML viewer plugin, but it was difficult to make the same SGML files work in both plugins. A standard was needed, and this was one of the incentives. W3C XML was born to meet the need of putting SGML documents on the Web.
We weren’t trying to force anyone else to use SGML (or XML), nor to do much more than make documents available on the Web. We succeeded in some of our goals, but in the meantime XML caught people’s imagination in ways we never predicted, some good and some not so good. We made friends and we made enemies. But now, thirty years after the birth of the Web, thirty-three years after the publication of the ISO SGML standard (ISO 8879:1986 SGML), twenty-one years after the publication of the XML standard, where are we?
If you have never heard a valley full of applauding dragons I must ask you to imagine rather than to remember. They start by clapping their front talons: a metallic, rasping noise. But then, as the excitement builds, there is a low roar, almost like the purring of a mighty cat. It quickly becomes deeper and louder and richer than the greatest organ-pipes and, as the dragons flap their wings and start to breathe fire, a sound so loud that no mere human could be near it.
With SGML and XML we built systems to to share information in ways that encouraged reuse. You can take an XML document and process it in any way you like: the format comes first, and the author, not the software. You describe what’s there and then build programs to process it. This is different from the Web, where the Web browsers call the shots. But Web developers are rediscovering some of these ideas with Custom Elements. In the meantime, XML is part of the World Wide Web, mature and solid and reliable. And we still fight for accessibility, inclusiveness, internationalization, and for declarative markup. For content, not glitter.
Happy birthday, World Wide Web, that has made such a difference to the world.
- All that is gold does not glitter,
- Not all those who wander are lost;
- The old that is strong does not wither,
- Deep roots are not reached by the frost.
[J. R. R. Tolkien]