XML is ten years old today. It feels like yesterday, or a lifetime. I wrote this that year (1998). It’s really long.

The title was originally Good Luck and Internet Plumbing but the filename was “XML-People” and I decided I liked that better. I never got around to publishing it, so why not now?

Remember, it’s ten years old; some of the people and companies are in different places now. [Update: In the comments, John Cowan writes Where Are They Now?.]


Ted · Theodore Holm Nelson looks like Peter O’Toole, with a personal style similarly extravagant. Ted’s Web page used to say “software designer”, but he’s never actually built any that mattered; in truth he’s a what-if person. He wrote two good books back in the sixties, Computer Lib and Literary Machines, asking, essentially, what if you put computers to work with language, not just numbers?

Put the right way, this question leads in some pretty entrancing directions; for Ted, they were the milk of paradise, and started him building a system named Xanadu. He’d invented the word “hypertext” (which is what both h’s stand for in http://www.god.com/love.html and similar billboard fodder everywhere), and Xanadu was going to be the universal hypertext that would contain everything worth caring about. A lot of people spent a lot of time and a lot of other people’s money and to make a really long story short, Xanadu never shipped. But it’s still a live project, Ted claims.

Whether or not Xanadu ever could have worked, or has a future, isn’t that interesting. Xanadu was one of the great what-if questions, and today’s World Wide Web is the best answer so far, which gives Ted his place in history.

He is for now in Japan, at a good university near Kamakura, where the great Buddha welcomes all comers. He’s had some tough years and he’s tired, but still strong, and still asking what-if questions. Mind you, not all of them are going to have answers as big as the Web. He’s a wonderful orator and I’d go out of my way any time to hear him hold forth on pretty well anything. Many people think Ted is wrong about any number of important technical issues, but asking the right questions is way more important than getting the right answers.

TimBL · Tim Berners-Lee is sometimes abbreviated TimBL, which rhymes with “thimble”. We have to call him TimBL because there’s another Tim in this story, namely me. TimBL’s a computer programmer by trade. He used to be a staff geek at CERN, the big physics factory in Switzerland, where hundreds of physicists from everywhere work in tribal shifts, spending tens of millions in hard-won research grants pursuing the neutrino’s mass (if any) and the Higgs boson’s existence (or not). These people produce information, lots of it and all the time; they need to share it around, but the printed scientific journals that used to serve this purpose are just impossibly slow and don’t help much.

By the time TimBL went to work at CERN, the Internet had been running for a few years, serving only big computers and the first wave of emailers. The CERN computer group wanted collaboration over the Net, and thought that hypertext, Ted’s invention, might be the way. So TimBL built a hypertext that worked across the Internet. Lots of computer scientists had been chiseling away at the problem for years, but TimBL ignored the entire body of theory and pasted together the simplest possible version with one-way pointers carrying no guarantees of what (if anything) they pointed at. No computer scientist could have conceived of anything so tenuous and fragile.

That turned out to be the World Wide Web, and TimBL is now famous. In mid-1998, the McArthur foundation recognized him as an official genius and wrote him a high-six-figures cheque, which is nice because a lot of other people who didn’t invent the Web have made a lot of money there.

TimBL is thin, pale, and twitchy, a well-bred British baby-boomer who circumlocutes and temporizes and gets to the point slowly. Englishly, he deplores confrontation and can find a way to paint any blood-feud in the colours of unfortunate misunderstanding. His publications suggest strong idealism, an overriding vision of the future of information space. His detractors say he’s a good second-rate programmer who was at the right place at the right time and got lucky. The McArthur foundation says he’s a genius. I can’t figure out what he’s getting at half the time, or why he does things, but I’ve known a couple of real geniuses and that’s not necessarily a symptom. As with most people in this story, you can read what he has to say about himself.

TimBL currently serves as Director of an organization called the World Wide Web Consortium (everyone says W3C), which is the Web family’s marriage counselor. By ’94, a lot of people had realized that there was some serious money to be made on the Internet, and a lot of little startups were entering the corporate toddler phase. The big parental figures like Microsoft and MCI and Sun had learned that none of them could (quite) corral the Internet for themselves, and it was obvious that their failure to get along might send the whole industry down the crapper.

So they built the W3C, and put a bunch of academics and dreamers and social liberals in charge of it. “For the children’s sake” so to speak. Which is not exactly in line with the ethos of late-modern capitalism, but it must have seemed like a good idea at the time.

If a bunch of businesses get together to agree on shared infrastructure, that arguably contravenes all sorts of antitrust laws. To get around this, the W3C is structured as a dictatorship, with TimBL having the last word on everything (the charter is reminiscent of Iran’s constitution), so that all the decisions and policies and standards are nominally the work of one person. And there’s no point in litigating against one person (even after the McArthur grant) for antitrust violations.

TimBL will certainly be remembered as the chief inventor of HTML (hypertext markup language) and HTTP (hypertext transfer protocol) which together are the glue that allows all the world’s computers to paste pretty web pages on the inside of your computer screen. TimBL now lives in Boston.

Ted and TimBL · Back when he was still trying to convince a skeptical public that the Web was useful, TimBL’s path crossed Ted’s at a conference; hypertext’s past meeting its future. Ted dismissed TimBL’s work as a toy, which was more or less true; but when when a toy’s range is the breadth and depth of the Internet, you have to remember that TV and the railways were also once toys.

It turns out that Bill Gates was also at that conference, and both Ted and TimBL were more interested in impressing him than each other. Both struck out.

Charles · Charles F. Goldfarb is a Jew from New York who got a Harvard law degree and went to work for IBM. How much more conventional can you get? This was in the sixties, and IBM hired him to help them sell their stuff to, and install it at, lawyers’ offices. Lawyers deal mostly in text, and at that time processing text with computers was mostly unexplored territory.

Charles and a couple of colleagues eventually invented something they called GML, which allegedly stood for Generalized Markup Language, but then Goldfarb’s two colleagues had last names beginning with M and L. GML turned out to be really useful in big-time text processing applications, and after some years’ evolution became an official “ISO Standard”, under the name SGML, for Standard Generalized Markup Language. Becoming such a standard is no mean feat. ISO, headquartered in Switzerland, is conservative, bureaucratic, and slow; irritating but arguably beneficial in an organization that exists to write rules for the world’s infrastructure.

Charles, fiftyish now, is dark, burly, balding, bearded, Californian. He wears Western-style string ties, talks energetically with a lot of hand movement and still a New York rasp. He’s been out of IBM and on the Web now for a few years, but remains synonymous with SGML. He has been known to say that developing standards is a war of attrition; the last person left sitting with unglazed eyes around the conference table often gets things their way. Charles has one of the world’s true iron butts; an argument with him is like trying to defend a sandcastle from an incoming tide. Furthermore, he’s often right.

This article isn’t really about technology, but I need to explain that SGML, despite the acronym, is not a language but a “meta-language” – you use it to build your own computer languages for your own purposes. Which is a very good idea, because the purposes of people are multifarious indeed, a custom language is often a good solution to a thorny problem. One of the most famous languages built this way is HTML – the one that TimBL invented for his working answer to Ted’s original question.

Yuri · Yuri Rubinsky died in early 1996 (suddenly, of a heart attack) before our story really got started, and his death affected a lot of people, including me; I excerpt here from my contribution to his collective online obituary:

I don't understand how a person can found a software company, tough it out through the bad days, lead it through the vale of Internet Fairy Dust, turn a silly idea into a PC Mag Editor’s Choice, simultaneously be a pusher, mover, and shaker in the WWW and SGML worlds, get a novel published, wrestle with ISO, and never, so far as I know, cut a throat, stab a back, or crush an ego. That it is possible to be successful and honourable at the same time is something we should take seriously to heart; in these days it seems too little respected, or even expected.

That alone would be a monument; maybe the only important one. But there are lots of others; a substantial part of the community, the technology, and the business.

[Recently …] he and I walked about 40 blocks home across San Fran …, late at night, talking excitedly about some stuff we want to do on the Web, that might yet turn into another monument. You know those hills, neither of us were exactly slim nor fit, and we laughed at our own heavy breathing as we slogged up, over and down. Tough to think that we won't be doing that again.

Yuri, a Canadian of Russian descent; was born in Beirut, Lebanon (a place we’ll hear from again) and, and while his family was Gentile as far as he knew, looked like a caricature of a mid-European Jewish intellectual. He came out of the typesetting business, built a new company and led it into the SGML business. He got Charles’ big book published and wrote its foreword.

Yuri was one of the first to see that TimBL was onto something, becoming an early Web evangelist, getting his company to ship one of the first useful Web authoring tools. Yuri believed passionately that while TimBL’s HTML was a cool application, the Web really needed something like SGML if it was going to grow up. But then he died.

Jon · Jon Bosak is a writer who escaped into high tech from the inspiring but stingy publishing world; not the first by any means. Almost any product has to come with instructions on how to make it go, and complicated products like airplanes and Internet switches and cars have to come with complicated maintenance manuals. This charmless necessary stuff is charmlessly called “techdoc”. Jon is a techdoc guy, one of the best.

He used to work for Novell, the Utah-based networking company. He did very well there, saving them tons of money and building one of the first really good and useful Web sites, where all the Novell customers could look up the latest and greatest techdoc without having to shuffle paper. This work was all based on SGML, Charles’ invention; Jon’s use of it to drive a Web site worked beautifully and should have been a signpost for the industry.

Jon, another boomer, is shortish, and looks like the child that might have been born to John Lennon and a Kennedy. His attire is resolutely and consistently brown and tweedy; he wears wire rims, and struggles constantly under the weight of loads of computer technology. We all carry computers now; Jon, however, has the biggest, heaviest laptop money can buy, and travels, just to be sure, also with a speakerphone and small printer. In his brown clothes, with bulging satchels, he looks like Kenneth Grahame’s Mr. Rat off to a picnic.

Jon is a highly effective platform speaker; he has mastered the unusual but good rhetorical technique of driving home a key point by lowering his voice, speaking more slowly, and closing his eyes.

Jon went from Novell to work for Sun Microsystems, a big, sexy, Silicon Valley computer company; he lives in Los Altos in the heart of the Valley. Jon and Yuri went way back, sharing the view that the Web needed SGML; but the Web was being built in suburban California warehouses by kids in shorts and T-shirts, who saw SGML as a big, corporate, blue-suit kind of thing. Nonetheless, Jon decided that TimBL’s W3C should make SGML happen on the Web. The W3C didn’t see it that way and ignored him; since he wouldn’t shut up and go away, they told him sure, he could launch an “SGML on the Web” activity if he did all the work, but to shut up and go away while he did it.

We’ll jump a little bit ahead in the story here; Jon’s SGML-on-the-Web eventually became known as “Extensible Markup Language” (XML for short) and is now a certifiably Big Deal on the Internet. So from here on in, we’ll just say “XML” and save space; but we’re really talking about Standard Generalized Markup Language on the World Wide Web.

James · James Clark is a computer programmer. People in the trade will tell you that a good programmer isn’t twice as good as an average one, but many times as good. There’s no meter to measure such things, but I’ve worked for, and with, and managed, a lot of them, and my experience is that the performance of working programmers varies by a factor of at least 100. James is at 100.

His ancestors made enough money that he, a GenXer, doesn’t really have to work. He is our story’s second Englishman, but is less polished and much more direct than TimBL. He has not let the fact that he is a deity in our technical community go to his head, is good company for dinner or a beer, and unpretentious on the Web. James is thin, darkish, open-faced, doesn’t stand out in a crowd; is erudite on wine, fine cooking, and text processing.

What James cares about is typography, and in particular how to arrange for computers to do it better than they do now. What he does about it is write fine software and give it away for free. James decided the world needed good SGML software. Unfortunately, SGML, as defined by Charles and his committee, was insanely complex and studded with options, many of which interacted in surprising and hard-to-understand ways. Thus, SGML software was scarce, expensive, and often flaky. James wrote a program called SP (for SGML Parser – one could imagine something sexier) that is so good that it has become the de facto definition of what is SGML and what isn’t. Naturally, he gives it away for free.

He lives in Bangkok because he likes Thai food and the language. He may have friends or family who are not Internet, programming, or typography geeks, but I’ve never met any. He works on problems that interest him. A German software company which once tried to buy some consulting from James (they subsequently retained me, which is how I know) were gently discouraged, with the explanation that he was quite expensive and really only interested in challenging problems. How expensive, they asked. If he were to come visit them and advise them on their problems, he would require a donation be made, in the amount of £10,000, to a Thai charity of James’ choice.

Jon and James · When Jon was launching the XML project, he took advantage of the W3C’s obliviousness to load the committee with people that he thought good, without regard to their corporate positioning. The first was James; since his name is one to conjure with, Jon persuaded him to accept the title of “Technical Lead”. He deserves his place in this story if only for having invented the acronym XML – if we assert that X stands for Extensible, then it does.

The XML project has run mostly by teleconference and email. James is not pushy on the phone, and his soft South England vowels don’t cut through the echo-y, static-y line from Bangkok very well, so we have to listen hard to hear what he says. Which we do.

Tim · This is me, a fortysomething Canadian raised mostly in Beirut (there it is again). I’m profoundly bookish and discovered early that writing is always difficult, sometimes boring, and usually poorly paid, while computer programming is mostly easy, only rarely boring, and remarkably lucrative. I’m a good programmer (somewhere in the middle of that scale where James scores 100), a competent writer, and (for my profession, where standards are low) an unusually good orator. This, despite my general disorganization, laziness, and lousy management skills, has been enough to get me through.

I’m average size, average weight, fair, largely bald, with a vaguely Mennonite look caused by wearing a beard sans mustache, and speak average toneless mid-North-American.

Before XML, the most fun I’d had at work was helping with the computerization of the Oxford English Dictionary. The OED, in its computerized form, was stored in almost-but-not-quite SGML, and I was a fan of the basic ideas starting about 1987.

A couple of colleagues and I started a company based on that technology, which proved only marginally interesting to the real world; we limped along for years until the Web came along and suddenly everybody wanted one. I built one of the first big Web search engines and we did an IPO and I got sick of corporate life and went independent. I live in Vancouver, Canada.

Jon and Tim · In May 1996, I delivered the closing speech at an SGML conference in Munich, exhorting SGML to simplify itself and get with the Web. Jon wandered up just afterward and asked me to join his XML project. I said no. A couple of weeks later I realized what a good idea he had, and called back to say yes. Jon said no; but then he changed his mind too.

Jon and I see eye to eye on most points, and appreciate each other’s obsessions. Also I think we work well together on a stage; Jon smaller, browner, still, focused; I larger, pinker, louder, animated.

Eve · Eve Maler, a Honolulu-born boomer now in suburban New England, toiled for years, like Jon, in the techdoc mines. I mentioned earlier that SGML (and XML) are used for inventing your own computer languages. Languages aren’t easy – birthing a new one is on the same order of difficulty as a screenplay, or a family of typefaces, or a housing development. Eve is perhaps the world’s leading expert on the practice of how you go about doing this; the book she co-wrote on the subject is probably the best reference.

Eve is buxom, curly-haired, full of smiles. Some people foolishly conclude that she is a technical lightweight and usually end up exposing their foolishness in public. In particular, Eve’s years in techdoc have made her a virtuoso among editors; no mistake, no matter how small, escapes her probing eye.

Michael · C. Michael Sperberg-McQueen has a name that is unusually long by any standard and was, during our many months of teleconferences, mangled every time by the operator. He knows a great deal about the use of computers to get complicated things printed, and is a language designer of some repute. Michael is erudite on a frightening variety of subjects, speaks perfect German, is a polished writer and a competent computer programmer. He became famous in the SGML community first as a gifted and stirring speaker – the perennial closing-keynote at our big conferences – and as a leader of the Text Encoding Initiative (TEI). The TEI is the result of years of collaborative work by many of the world’s best humanities-computing academics. It is a family of SGML-based languages which are used to store an amazing variety of electronic scholarly texts including Chaucer, Strindberg, the correspondence of Proust, Serbian Proverbs, Voltaire, and Japanese Map Task Dialogues. Michael, about 8 feet tall, correspondingly broad-shouldered, and with a bit of middle-aged spread, would be an intimidating presence were he not so professorial. He is at all times in strict conformance with the dress code of the Midwestern academic: tweeds, checks, scarf, cloth cap.

The Honeymoon · I guess at some point I have to mix some events into this story. Jon formed the committee, which started work in August 1996. Michael and I were appointed co-editors of XML. We worked out the technology in telephone conferences and group email. On the committee were some other people that don’t get their own write-ups here; this is totally unfair, because they did as much of the work as the people you are meeting. But since I’m the writer, I get to pick the cast of characters, based on how much fun they are to write about.

In one sentence: XML is a meta-language built by taking SGML, throwing out most of the little-used advanced features, building in good foreign-character support, and a few other technical tweaks not worth expounding on here. The world seems to like it.

Our first draft specification came out in November 1996. Its construction was a wonderful experience. Jon had stacked the committee with people who knew meta-languages, who understood the Web, and who knew what worked and what didn’t. Jon did not invite Charles on board, and whereas Charles has never said a word about this, it is reasonable to suppose that he is hurt to this day. We all knew each other, used the same vocabulary, and wanted more or less the same thing. We had technical differences, passionate at times; but the progress was fast and steady.

Charles and Michael and Tim · We on the committee were working hard, designing XML at top speed and having fun doing it. Eventually, we reached the vexatious question of whether we had to be 100% compatible with SGML; Michael and I had some ideas for changes that felt like real improvements. Charles got wind of this and went nuts. To introduce our changes would have been to betray our heritage, to split the industry, to throw away a decade’s gains, to spit, it seemed in the face of all that was good and sacred. So he launched a war of attrition, his chief tactic being to telephone the members of the committee, one by one, and fill the phone line with heated (and in some part reasonable) arguments against our apostasy. He was not unreasonable, he was not sleazy, and he was not unfriendly. He was also not willing to get off the phone until you’d conceded his point. Since he had a habit of calling up after 8 PM, this could get seriously trying.

Charles’ butt proved more ironic than those of a substantial number of committee members, his incoming tide washed away the opposing arguments. Enough, eventually, that the committee was split on the issue, more or less evenly, and most of the new ideas died unborn (in retrospect, this was not entirely a bad thing). It still seems unreasonable to me that one person can by sheer endurance beat down a committee of 11, who are on average 10 years or more younger, but there you go.

Lauren · Lauren Wood is a New Zealander whose adult life has been in Australia and Germany. She has a Ph.D. in theoretical nuclear physics and speaks fluent German, but discovered the same work-pay relationship between physics and programming that I discovered between writing and programming. She has now been toiling in the publishing software mines for years.

Lauren is smallish, slenderish, dark-haired, pale-skinned, large-eyed, and speaks with an Auckland lilt that can veer into complete opacity on words with a strong “O”, such as “word” and “cold”. She is cheerful, patient, and very good company.

Tim and Lauren and Yuri · I met Lauren at a conference in Montreux; she was working for a German company that sold, among other things, my company’s software. Then I met her again at a conference in Singapore, and she visited me in Vancouver. Pretty soon I was trying to convince her to stay in Vancouver, but there was the little matter of a job and citizenship and so on.

Fortunately, I ran into Yuri at an airport. In the near-empty pallid grey of the baggage-claim, he twisted my arm to help him write a book about SGML on the Web. I twisted his arm to help get Lauren a job in Vancouver (where he had a programming shop). Since I was short of time, he failed to get my help with the book. Since our industry is short of talent, the job was no problem. Canadian immigration issues melted away like the infrequent Vancouver snow in the face of a native English speaker with a Ph.D., a marketable skill, and a company begging to hire her.

In her new job, Lauren quickly got involved in the XML project. We were married in November 1996, just after XML was unleashed on the world.

Jean · Jean Paoli is a citizen of France but really Lebanese, a typical Levantine ethnic cocktail: dark, burly, hirsute, chubby-faced. Jean, like I, left Beirut for more favorable climes when the people of Lebanon, funded and encouraged their neighbors and by the great powers, took up civil war as a pastime. He was the main brain at a little French software company that sucked at the teat of the European Union and other continental bureaucracies, and had a nice line of SGML products very early on.

Jean was closely involved with the European research establishment, and clued into the Web just about on day one. His was a famous face in the SGML world, and everyone noticed when, one day, he packed his bags and moved from Paris to Redmond to work for Microsoft.

Jean was a charter member of Jon’s XML project; the idea that Microsoft would actually pay attention to our dream was more than a little intoxicating. Jean, a leading co-conspirator during XML’s honeymoon, soon leaves the story, for reasons we’ll get into later.

Phil · Phil Karlton died in 1997 (in an Italian auto accident while on vacation), so only takes a brief turn on this stage. Way into the top half of that 1-to-100 programming proficiency scale, he had a good academic career, then worked at the revered Xerox Palo Alto Research Center. Xerox PARC was where they invented icons and pointers and windows and workstations; but management was too dumb to ever make serious money from these things.

Phil went to work for Netscape, the startup darling of the Internet, not in its first days, but around the time it achieved critical mass. His job there was to provide general technical leadership; he was old enough that his son was also on the Netscape programming staff.

Phil was an ugly bald whiskery sweet man, with lots of rough edges and an apparently-perfect 20-year marriage that, I think, intimidated many whose relationships are more mundane and compromise-laden. He drove an NSX and played ball-hockey and lived intensely; all this is well-documented.

Lauren and Tim and Phil · Phil, his wife Jan, Lauren and I are meeting for dinner at a nice Italian restaurant on El Camino in the Valley. Since we have a positive personal buzz, the deal is that this is a social engagement, not a technology debate nor a plot-hatching session. The hostess leads us to a small table beside a large, boisterous party. Phil asks her who the people are (you need to be careful in the Valley). From Microsoft, she says, and Phil, in his booming rasp, answers without a pause “Microsoft? I won’t sit beside them – let’s go over there.”

Over dinner, despite our best efforts, conversation veers again and again into matters of Internet politics and markup languages and software technology. Phil lets this happen, but every time a new acronym or piece of jargon or person’s name flies across the table, he stops, turns to Jan (who is an artist – not geeky at all), and explains in clear layman’s terms what or where or who it is. Such consideration is not common; even at second hand, it is moving.

Ned and Mick · Some of the people in this story are companies. Ned is Netscape and Mick is Microsoft. Yes, any company, au fond, is just a collection of people. But some companies, at some times, are more than that, and their representatives are both less and more than people. Less, because they perforce speak with a voice not their own, and more, because that voice is the voice of hundreds or thousands (of colleagues), and of millions or billions (of dollars).

For example, while Jean was just Jean at the beginning of the process, once Microsoft got really interested he stopped being Jean and starting speaking with Mick’s voice. It’s hard to argue with a company, because they make their decision off stage, and their role in a dialogue is then to win their point, not to strive for truth. And as XML became increasingly successful, Jean’s role on our committee was filled by a succession of other Microsoft staffers, whose individual personalities matter little.

Mick is a domineering, ruthless, greedy, egotistical, self-centered, paranoid bastard. Whether or not he’s actually a crook is, as they say, currently the subject of litigation; but he’s not good company or a good friend. The ruthlessness and greed would not be so irritating (we swim, after all, in late-capitalist waters) were they not accompanied, at all times, by Mick’s claim to speak not in his own interest, but selflessly on behalf of his millions of customers, whose needs only he understands. Thus, anyone who disagrees is conspiring against the interests of the world’s computer users.

Mick’s other really irritating habit is constant grating prating about “great” products and “innovation”. Certain Microsoft executives are going to spend eternity fleeing around the bolgias of Hell from demons wielding branding irons on which “great software” and “innovation” glow white-hot. A very large majority in the computing trades think the products are mostly pretty poor, and see the company as the single greatest roadblock to innovation in our profession.

Netscape a.k.a. Ned was definitely the new kid on the block. His parents were an academic staff geek who added pictures to the Web, and a money manager who’d launched a couple of high tech high flyers. Ned came from nowhere at just the right time and suddenly Owned The Internet, in large part because Mick hadn’t noticed it; if Bill had paid more attention to Ted and TimBL at that long-past conference, the world would be different.

Coming from nowhere and Owning The Internet turn out to be bad influences in the young life of a company. Ned was like a rookie pitcher who burns ‘em up in his first 20 games with a wicked screwball, but won’t listen to the pitching coach who tells him to mix them up, nor to the trainer who tells him he’ll burn out his arm, nor to the manager who tells him the hitters will eventually figure that screwball out.

While Mick got with the XML program, Ned steadfastly ignored it. Jon’s XML gang disliked this, so all of us pounded on Ned’s door, besieged him with email, called everyone we knew with a line in, trying to send the message that this was important, and that Mick was getting into it big time. Ned didn’t have time for us, because Ned Owned the Internet and don’t you forget it.

Lauren and Phil · Lauren had met Phil in connection with another W3C activity, and they had hit it off. She was known to be one of the SGML bigots who were playing this XML game, and yet she seemed young, Web-oriented, pragmatic; not one of the crusty ISOfied bureaucrats.

So Lauren got a call one day, and took a quick trip from Vancouver to Silicon Valley. It’s the same time zone, you can get up in Vancouver and make a 10:30 meeting down there, which is fine because technical meetings in the Valley are never before noon. She spent the day closeted with Phil and a bunch of other senior Netscape hacks, and, it seems, made the sale. Ned was now officially an XML fan.

Ned and Tim · Trouble was, Ned still Owned The Internet, except for Mick was pouring hundreds of millions into Internet R&D so he could give software away, drive Ned into out of the business, and Own The Internet himself. Thus Ned was preoccupied, and didn’t really have anybody on staff who understood XML issues.

The solution was pretty obvious to me. I was independent, remember, and I’d been doing all this XML work pro bono, first of all because it was a good thing to do, and because I thought I’d have some spare time while my business picked up. It turned out to be hot stuff, and very time-consuming, and I didn’t want to let go, but I did want to get paid.

So I gave Ned the big pitch – offering myself as their eyes and ears; they’d get to find out what was going on, they’d visibly be no longer ignoring XML, and I’d make some money. They bought that pitch, which was certainly a good deal for me; basically I was getting paid for what I’d been doing anyhow; all I had to do was to write them a weekly email report, and pay regular visits to brief their engineers.

Except for those regular visits never happened; their engineers were too busy Owning The Internet to pay attention to outside ideas.

Eve and TimBL · It’s early 1997; at a nice restaurant in Santa Clara, a few of us gather for dinner with TimBL. This is the first time he has actually spoken with the XML gang, and the first time some of them have met him. I am sitting next to Eve, across the table from TimBL.

TimBL starts to expound another of his brilliant ideas, which while technically not half bad, exhibits profound ignorance of a decade of history. Imagine a young priest sitting down with a conclave of cardinals and explaining that the Catholic message would be so much more coherent if they dropped all that rococo blessed-virgin-Mary-mother-of-God stuff. After a certain amount of this, Eve is literally shaking with fury, the volume is up, and the waiters are watching our table nervously.

On the way back to our hotel, Lauren, looking puzzled, asks me a reasonable question: “were you flirting with Eve?” And I am forced to admit that I had put my hand on her thigh more than once. But these were cautionary cool-it-buddy pats; I’d had this vision of Eve hurtling across the table to get a stranglehold on TimBL, and I’d bet on Eve in that fistfight any day. It would have been seriously bad publicity for XML if one of our committee members were publicly to inflict violence on the Inventor Of The Web.

Mick and TimBL and Tim · First thing I did in co-operation with Ned was publish a paper that got a lot of attention around the W3C and sent some smiles Ned’s way. Mick went ballistic; this was the period before the Department of Justice had really started nosing around, when Netscape delenda est was the tune of the day in Redmond, and no blow was too low.

So Jean called me up and wondered if I would drop Ned and start working with Mick; the details didn’t really matter, but the price would be good. I declined, since I was having too much fun, and as I told Jean, Ned really needed the help. “Tim,” said Jean, “personne n’est irremplaceable,” but his voice was sad, because I wasn’t buying it.

So Mick called up TimBL and said “one of the XML co-editors has been co-opted by the evil Ned, and we can’t have that”. I was ordered to resign as co-editor, and Jon was ordered not to reappoint me without unanimous consent, which Mick would of course withhold.

I refused to go gentle into that good night; in fact, I sustained a childish tantrum for some weeks, scurrying around soliciting support from other committee members and publishing an intemperate Philippic. Mick’s idea was that my defenestration should be accomplished quietly and with the minimum of fuss. Thus my shrill public squealing was acutely embarrassing for all concerned, and gave the Mick-haters (there are lots) opportunities for gleeful if not entirely sincere Expressions Of Grave Concern At This Questionable Process.

In all this nastiness, Jon distinguished himself by (outward) reasonable calm, and by refusing to do anything on grounds of expedience, requiring rather a basis in principle for each action. Three weeks after my execution, I was exhumed and revivified; you’ll still see my name on the front of the XML specification. It probably helped that it would have been difficult at that time to find someone else with the spare cycles and expertise. As part of the deal, Jean was named the third XML co-editor, which was a good idea anyhow.

Nov. 14, 1996 · XML was officially unleashed on the world on Nov. 14, 1996 at the annual SGML conference, which draws a thousand or two to the frozen Northeast. Nine members of the XML committee sat in a semicircle on the stage. Jon and Michael and James and I spoke to a packed and electric room. James, who has a tendency to mumbling and jargon, was dynamic, even cracked a joke. Charles was not in the audience. The SGML community, smelling Internet money and the mainstream, took about 15 seconds to buy in.

Jon and I turned our attention to selling it to the rest of the world. We besieged technical journalists, scrambled for speaking spots at every significant industry conference. We thought we’d face a brick wall of resistance and indifference and were determined to batter through it.

It melted. It turned out that Jon and Yuri had been right all along; the Web apparently needed more or less exactly what we’d been building. You can read about it now in newsstand magazines – TIME and The Economist, forsooth – and a few of us, Jon and I as chief evangelists, but others as well – have become low-grade Internet celebrities as a result.

Of course, then the honeymoon was over. The people in this story don’t matter any more. The voices around the table – and there are more of them – are those of Microsoft, Sun, IBM, Adobe, and so on, not of individuals. There’s less debate and more posturing. Some voices are lightweights and add little value, but each jealously demands its time slot.

This isn’t about technology any more, and certainly not people, it’s business. The Internet business, for all the visionary rhetoric, has to do with nothing but money and power and executive ego.

Dinner With Ted · Spring 1998 in Brisbane, Australia. This is the 1998 version of the same conference where, in 1996, the XML idea was born. This is the official World Wide Web conference, and Ted has been invited to receive an award for his contributions; in his remarks, he hushed the thousand-odd people in the room, saying “Thank you. This is a first; nobody’s ever given me an award before.”

That evening, someone organizes a dinner date for Ted and a bunch of the XML tribe, including Jon and Eve and me. It’s a lovely restaurant on the riverfront with good food and good wine and lots of it (Australia’s like that). A warm wind blows in scents of the Great Barrier reef. We talk and talk about fine books and fine computer programs and fine food, and we disagree often, at high volume and in high good humor.

It’s all luck, it’s just luck sings Joni Mitchell, and we’re in a profession and at a time where there’s a lot of it about; still, sharing laughs around the table, on the waterfront under Southern stars, with an inventor of the world you work in, that’s really a lot of luck.

James and Lauren and Tim · Spring 1998, late evening in a hotel lobby in Seattle. We’re at the first-ever big commercial XML conference; though poorly marketed and programmed, it’s drawn over seven hundred people out of the woodwork. Clearly we’re catching a good break in the Internet wave. James, true to form, has hidden for a few weeks in Bangkok and quietly built some XML software that is the best and fastest anywhere. From the stage on the morrow it will be announced that it’s getting wired into two of the Internet’s most influential software suites (Mozilla and Perl, for those who know such names). We’ve realized, over dinner, that James’ working name for his code, xtok (sounds like something you buy at a head shop), just won’t do.

We sip a good local beer and have fun with names… James is settling on expat, which might stand for “excellent XML processor and toolkit”, on top of which he is one. Then a bloated suited Industry Leader sees us inexplicably ignoring all the other suited Decision Makers drawn northwest by the smell of money rising over a new acronym. He butts in to offer his opinion on Strategic Media Convergence or some such, so we finish our beers and go to bed.



Contributions

Comment feed for ongoing:Comments feed

From: Eric Meyer (Feb 10 2008, at 11:36)

Thanks for posting this, Tim. All too often the people and currents get lost in time, and it's always important to remember the human dimensions behind the technical achievements.

Ah, Brisbane. I think the same evening you were having that dinner, a bunch of CSS geeks were having their own mass dinner somewhere nearby. I'll have to see if I can dig a photo out of my paperstock archives.

That's also the conference where you and I first met, albeit incredibly briefly, Tim. Not that it matters to the story, but it's fun connecting the dots of other stories to this one.

[link]

From: John Cowan (Feb 10 2008, at 11:40)

Where Are They Now?

(which used to be the title of a New Yorker department in the early days)

Ted is a visiting fellow at Oxford, and Xanadu is open source, for all anyone actually gives a hoot.

TimBL is still at W3C, and has been knighted by his Queen.

Charles is in California doing the independent consultant thing.

Jon is trying to get the flow of documents (invoices, POs, and so on) that keep business in business properly electronified, XMLized, open, and free.

James is still in Thailand, but if anyone knows what he's actually doing, they're not talking.

Tim is working for Sun as Lord High Fixer, and running this blog. He and Lauren are still together (heading for that 20-year figure) and have reproduced repeatedly.

Eve is working for Sun too, with an emphasis on identity. This is somehow connected with knitting, but nobody else understands exactly how.

Michael is working for the W3C, preparing for the second bite into the uglifruit that is XML Schema.

Lauren is also at Sun (seeing a pattern here?), also in the identity business, having survived the attack of the crazed weasels as well as the five-year role of tentpole in the XML big tent. See also Tim.

Jean is still with Mick, and is now regularly addressing the world in the first person plural.

Ned is now extremely deceased (after years of being just mostly deceased), but his legacy will endure forever, or as near to forever as things get in this business. This annoys Mick no end.

Mick still bestrides the world like a Colossus, but "the little rift within the lute / That by and by will make the music mute" is not so little any more.

Yours truly, who doesn't belong in this post because he had never heard of XML or SGML or anything markup-related at the time, and whose first XML document had no quotes around the attribute values (ptui!) and existed only on a patent application (double ptui!), is hacking Atom and APP for Google.

[link]

From: Keith Fahlgren (Feb 10 2008, at 11:41)

That was great, Tim. Thanks for (re)publishing it.

[link]

From: Stéphane Lacelle (Feb 10 2008, at 14:59)

I hate doing this on such a good paper, but it should have been “personne n’est irremplaçable ,”...

[link]

From: John Cowan (Feb 10 2008, at 18:15)

Misspelling our French is part of the anglophone tradition both within and outwith Canada. Consider Le Morte Darthur, or "le utility del chose excusera le noisomeness del stink", which is why a soap-factory is not a nuisance at common law. Or how it is that the English parson (just "person" in another spelling) is a vicar and his assistant a curate, whereas the French priest is a curé and his assistant a vicaire!

[link]

From: Hugh Barnes (Feb 11 2008, at 04:51)

Thanks - that was exceptionally well written, engaging, and an enriching way to spend the best part of an hour.

I have a teeny contribution of geographic pedantry. I think the only Barrier Reef olifacts you'd be getting from a Brisvegas riverside restaurant would be wafting from a nearby kitchen. The Reef starts hundreds of kilometres to the north. Fire your continuity guy. (… and sorry to spoil your romantic image, I've been waiting to use "wafting" in a sentence.)

[link]

From: Alexander Falk (Feb 11 2008, at 07:13)

Tim:

thanks for sharing this - great story! I originally wanted to write a 10 years of XML blog post myself, but then decided to just link to yours instead. There isn't much to be added....

Cheers,

Alexander

[link]

From: Omar Gonzalez (Feb 11 2008, at 07:15)

That was a great read! At the time this was happening I was just a freshman in college, without a clue that I would be making my living 10 years later on this new thing called the world wide web. Great stories!

[link]

From: Justin Watt (Feb 11 2008, at 10:08)

I'm waiting for the Flash movie

http://www.youtube.com/watch?v=2pPCkhYMQgY

[link]

From: Edward Vielmetti (Feb 11 2008, at 11:57)

In September of 1990 I created the Usenet newsgroup comp.text.sgml

http://groups.google.com/group/comp.text.sgml/browse_frm/month/1990-09

by going through the Usenet create-a-group process. The group immediately had useful and interesting traffic, with all of (most of?) the above mentioned people involved in it.

[link]

From: Derek K. Miller (Feb 11 2008, at 22:09)

Sorry, I'm a bit confused. The piece says "XML was officially unleashed on the world on Nov. 14, 1996 at the annual SGML conference," but back at the beginning you wrote, "XML is ten years old today." So is it ten, or eleven and a half?

[link]

From: Tim (Feb 11 2008, at 23:19)

Ed: I poked into comp.text.sgml and look what I found.

Derek: The first public draft was November '96, the final release as a 1.0 was Feb '98.

[link]

From: John Turnbull (Feb 14 2008, at 11:41)

Thanks for this Tim.

Personal histories are so ... personal and I am reflecting on what very lucky guy I am to have passed, like a small meteorite, through this planetary system of XML inventors. I met many of the personalities you described so well and others who had contributions to make.

Thanks to all of you, and particularly Yuri, who, in addition to all the frantic activity you listed, had time to offer a neophyte "electronic publisher" some intriguing advice.

[link]

From: Chris Maden (Feb 16 2008, at 20:03)

Thanks for writing this up, Tim. Good times, they were, and I still have the (now-syntactically-invalid) <?XML!> lapel pin at work. It was a joy just to stand on the sidelines and occasionally catch ball headed out of bounds.

[link]

From: craig smith (Feb 18 2008, at 00:49)

thanks for the great article, i am nby no means old at only 22 but i always remember, even before the final specification, my dad getting involved. he had been in charge of a company who developed an sgml search and retriev engine. small and fast. they then got to work on changing the engine to support xml, and it too must be nearly 10. at the end of the day, xml is not being used where it is sorely needed, in the heart of data storage and retrieval, those that try to implement it feel that is good to bolt it onto the top of an archaic system, like say, oracle! you guys have the answer to a huge bulk of problems, but large businesses arent putting your hard work to good use as far as i'm concerned. again thanks for the great article. big love from craig, scotland, on the bus home.

[link]

From: Iain Craig (Feb 18 2008, at 08:44)

Great article, Tim.

"Xanadu was one of the great what-if questions, and today’s World Wide Web is the best answer so far, which gives Ted his place in history."

- I wonder if the Wiki engine, and by extension Wikipedia, is a more recent answer to the same question.

[link]

From: Greg Fenton (Feb 18 2008, at 21:22)

Doesn't Larry Wall work into this story in some capacity? Or is that an entirely separate off-shoot?

[link]

From: Mike McNamara (Feb 19 2008, at 01:02)

10 years old! Very interesting to see some of the history and the various people involved over the years.

Although not 'involved' at the centre of XML (or SGML), it is interesting how 'mark-up' and its use has developed since my first involvement with it through Xyvision Inc. which I joined back in 1985!

I think that the next 10 years will be as equally interesting as the past 10 years have been.

Mike

[link]

From: Alan (Feb 19 2008, at 01:58)

very nice! thanks!

[link]

From: Sandy Ressler (Feb 19 2008, at 09:39)

So good to see Yuri's name again. Twas an amazing guy...one of my personal favorites of his was "SGML The Movie" a wacky, yet informative little short introducing many of SGML's concepts.

[link]

From: Richard Hildred (Feb 19 2008, at 10:04)

This is a wonderful article. I had an interview with the Oxford dictionary people in Waterloo in 1994, but decided to stay on as an employee with the large US company that I was contracting with in Waterloo.

Your people seem to have had a lot more fun, even though I also love xml, expat, xsltproc from their infancy, and can code my way out of a paper bag with them.

thanks again for the great article

Rich

[link]

From: Al (Feb 19 2008, at 17:59)

Why was I left out?

[link]

From: ju (Feb 22 2008, at 09:10)

so nice! thanks a lot!

[link]

author · Dad · software · colophon · rights
picture of the day
February 10, 2008
· Technology (77 fragments)
· · XML (135 more)

By .

I am an employee
of Amazon.com, but
the opinions expressed here
are my own, and no other party
necessarily agrees with them.

A full disclosure of my
professional interests is
on the author page.