Perhaps the most telling thing about my experiences at Extreme is that when I was asked to write about them, I settled on the papers I had missed, misevaluated, misunderstood or daydreamed through.
I can tell you now that I captured virtually every word that Matthijs Breebaart of the Dutch Tax and Customs Administration said this morning — because what he was saying had pertinence to what I am doing, because it will be useful for me to pass this on to people I work for and with, because that way what he said would stick in my mind better, thus cementing my understanding and helping me when I sat down to write this essay. Then I got his slides and the written paper. I don't want to miss or mistake any of this.
Was Matthijs' presentation the most important to me? Three-quarters of the way through the day's program, I have to say Yes and Who knows? The reason for my ambivalence is because of a little thing Wendell Piez said in his colloquium — oh, it was a colloquium not a presentation, even though the audience held its comments to the end: there were no questions, just "now let me add to this too" comments. When I first began to teach, it was in a poetry-in-the-schools program for third graders and I learned that the teacher has successfully imbued his students with the right notion — with education, with ideas — when the kids are bursting with questions and want to turn you off and head out on their own and do something with what you had just ignited. (And no ignition, no learning, btw.) And that's what this session with Wendell was like, in which he confessed almost deliriously that he had been sinning against the basic precept of markup, that is
As it happens, I come out of a completely different school of publishing from Wendell, one in which there was no one called a markup designer, in which presentation occupied more effort than any other aspect of the creation of the page. I've been a closet sinner for some time. Where I worked, designers ruled. You could give a designer a picture of a person and the two words Oh, yeah? and he or she would gladly spend all day arranging the two items at different sizes and in different colors and in different fonts (for the text) and trying uncounted things with that question mark until the words could be removed or randomized with Elliotte Rusty Harold's obscurer tool and you would still be able to tell what that page was about. That was what designers considered the ideal — the semantics of the article could be deduced from what they did if all the editors and typesetters managed to substitute in text from The New York Times classifieds. That every part of the p
So, anyway, in the midst of his confession of heresy, Wendell connected what he was discussing with what Walter Perry had been talking about before lunch. Well, there's practical and there's abstract, and Walter falls so far on the side of abstract that I was reading his slides as though they were graphics and disregarding the words. (Of course, that's my upbringing.) Anything to stay in that moment. And here in the midst of my internal combustion listening to Wendell to have this flame of brilliance connected to the haze I'd earlier been in was pretty jarring. As I've noted elsewhere, more than once I've found the least useful papers I've encountered at Extreme proved unerringly to point to the path I would be traveling in a couple or three years. And Wendell was connecting tab A to slot B right there.
So maybe what I should be thinking about is Walter's talk. However, only one part of it stuck in my mind, when he referenced "Finnegans Wake vs Burger King Pickup Ticket". I thought, "Whatever the connection, the juxtaposition of these two types of document is stimulating." That's all I thought, however. Fortunately, when asked a question by C.M Sperberg-McQueen, Walter supplied a remark destined to be widely quoted: "I am a great believer in the bogus-ity of authorial intent." (I guess it is right that we hold our conference chairs to high standards, that they elicit such remarks with regularity.)Wendell's seminar had, of course, the side benefit of exploring his SVG-based presentation tool, with the woolly fashion in which its pieces are assembled. To look at the full presentation reduced to a single view is like looking at one of those maps of an art museum with rooms and paintings and map features all laced together. Bracing, really, and yet almost dreamlike when compared to the ruthless march of bullets in every other presentation [1]. More than anything else at this conference, it reminds me that we are dealing with computing devices of overwhelming power, and that that power can be harnessed in ways other than speeding up ever-larger processing tasks. Of course, Wendell's talk has the same effect, with its references to McLuhan, sonnets, Dewey, Leonardo, Dali, as subject matter but also as conjurors of nonrepresentable information (can I call that knowledge then?).
When Matthijs spoke in the morning's third session, I may have fallen victim to the standard conference imperative: Show me something I can use when I get back home. A method for dealing with multiple copies of the same document, structured differently and under the control of different groups of people (or vendors) who are responsible for its almost pathological updating — we're talking legislation here — yes, that would be good to have. A method for dealing with time validity. An approach to constructing identifiers, with a serious pitfall or two to avoid. Discussion for reasons to choose meaningful identifiers over meaningless ones with a lookup table, and an explanation for his choosing a composite identifier, with both a meaningless segment and meaningful ones. Why the identifier was built using Relax NG with XML Schema datatypes in order to be self-describing, and how XSLT was used to transform it into a simpler-to-handle URI.
Not to put too fine a point on it, but I can see how every part of his experience might benefit the publishing company where I work. That's direct benefit, at least to me. But I think I was drawn in more by the real-world quandaries he faced, and as a check on how he acted in situations we face too. We have two needs pulling us in different directions, he said. Things are so expensive, we don't want to pay for any material that's now available for free. And we must always present every scrap of information that might be useful. He took, almost at random, a single article of the Dutch tax code and showed four different versions from his site — with identical text but different structure and organization, and all with different URIs — from four different vendors. We have a huge amount of information, he said, but it's organized in vendor silos, and you have to choose the silo you want first and go from there. How could the vendors be brought t
OK, case studies you get at every conference. But not ones where Topic Maps are integral to the system being built and yet only make it into one slide because the heart of the matter lies in the unexpected use of XML — the old technology at an edge markup conference — for a self-describing identifier. I liked it. There was so much to chew on; opportunities for misadventure successfully sidestepped. A good day at Extreme, the kind you always dream about having at a conference, with practical juxtaposed with abstract, the obviously thought-provoking linked inextricably with the seemingly better forgotten, Joyce cited and McLuhan studied. Thank you, Wendell. Thank you, Walter. Thank you, Matthijs.
— Roger Sperberg
So that you can see what Matthijs was talking about, here is how the identifier in XML looks (translated from the original Dutch):
An example identifier. Courtesy Matthijs Breebaart.
1. Except for Ann Wrightson, who drew pictures and handwrote her text in colored marker.