May 08, 2003
XML Europe conference report (long)
Sitting on the Eurostar back home, let's quickly jot down a report on
what has been a great edition of XML
Europe. While the conference organization and content might be
subject for some modernization, I had a terrific time meeting some very
interesting people. It's funny to see, even for a regular attendee of
this series of conferences (it must have been my eight edition or so),
what difference the black 'speaker' ribbon does make, which I partly
blame on this weblog. :-)Upon arriving at London after an uneventful journey (besides learning there exist two Edgwares in London, which cost me a fair amount of unnecessary underground tickets), and after checking in, I went to see how Marc was doing teaching a Java/XML/Open Source full-day tutorial. That evening, the conference organizers hosted a reception to thank the speakers, so off we went. We met two interesting guys over there, Jacek Ambroziak who is the author of Gregor, the (even) speedier successor of XSLTC (for which he was the original author), and Farrukh Najmi, working for Sun from his basement, and the lead guy on the reference implementation of ebxml/rr, an open source ebXML registry and repository. Even when we were all slightly intoxicated due to the amount of French (!) red wine we got to consume, Farrukh managed to widen and deepen my knowledge about the goal and scope of ebxml, which clearly is not restricted to business documents and transactions only. Farrukh was a guy with great passion on the standards and RI he was working on, and it was a joy meeting him.
Meeting Jacek was quite special. I introduced myself as one of the guys who nagged him over a year ago about recontributing his fork of XSLTC (now called Gregor) to the Xalan/XSLTC project, together with Stefano and presumably some others. Jacek also knew me from distributing an XXE setup for the conference paper XML grammar, and both Jacek, Marc and I hit off on shared neurons big time. We spent many hours together in the next two days, we managed to pick his brains on the implementation details of Gregor (and consequently also the reason why he manages to be significantly faster than any other XSLT engine out there), a process which required him to open up his brains in a very confrontational way, up to the level of showing Gregor's source code. Jacek, besides being a great guy, is one of these software engineers who happens to think genuinely out-of-the-box, and we really hit some common nerves which made his initial caution in letting us in on Gregor's implementation details quickly fade away. As I said to him upon leaving, our encounter alone was worthwhile the journey. It also learned me how influential our nagging had been upon Jacek's decision to seriously continue work on Gregor after our vein attempts to have him reintegrate his XSLTC patches with Xalan for Cocoon's purposes. Which taught me again how influential Stefano's vision and way to work with people can be - kudos to him for that. While Gregor might not be open sourced soonish, I have come to respect Jacek's decision, now that I know his out-of-the-box approach and the effects this approach would have, rippling down through the entire Java XML pipelining architectures out there. Think SAX-NG for that matter.
After a bad night of sleep (a stupid habit of myself to sleep really bad on a first night in a foreign bed), I went to see the opening keynotes of Jon Bosak and Daniel Veillard. Jon, who is the guy who brought the team which created XML together, invited us to take a look at UBL (an invitation I might pass on this since I don't necessarily believe in all-encompassing vocabularies, especially if part of them are strongly inspired by so-called existing implementations (OpenOffice) rather than the implementations tracking a clean-room standard. But then again, there's some other (JCP) standardization tracks which basically endorse a commercial product's API as the 'reference', a trend which is worrying me these days, which leaves me pretty much in doubt whether to adopt JSR 170 for our (hopefully) upcoming content management project.)
Daniel Veillard was supposed to give an overview of the usage of XML in the Open Source community, but suffice to say that he managed to really piss me off with the tunnel vision of the FSF and consequently RedHat. He skipped the entire load of Java/XML Apache projects because of lack of time (fine with me), but also because they were running on a non-free platform (Java). Yeah, you heard me right on this one. I start to think the real problem is that of RedHat not being able to ship Java with its free OS without paying Sun (which sucks). Having the end-user install a Java runtime however is a real piece of cake nowadays - so condemning Java/XML Apache projects on the fact they require a non-free prerequisite is like a harsh way of saying "we don't like Java just because we don't". Which leads me to believe that RedHat wants to have the same level of control on what is packaged with your desktop OS as that company of Redmond does. Anyway, we met and discussed and I now know for sure that is prefer Freedom as a Right rather than a Duty.
Over lunch, I sat next to Dave Pawson, who I met many years ago on what is supposed to be the only public European DSSSL course ever, and which since then has become an XSL-FO authority, being an O'Reilly book author and all. We shared thoughts on organizing grassroots conferences on emerging topics (he did for XSLT, I did for Cocoon), and about becoming an author. I might.
Then, I had to chair the Publishing track with two more print-oriented subjects being brought forward, and after that David Megginson showed up, stating he would be chairing and introducing my Cocoon presentation. Which was way cool: the original author of SAX who comes to introduce me, talking about a SAX-centric framework. The presentation went well, with some very familiar names in the audience: Jeni, Daniel, Jacek, Simon, Uche and Paul (I believe Edd dropped by, but am not too sure about that). How's that for an audience to be really nervous about? Seriously, it was a real adrenaline shot to be presenting in front of these people, and after the talk, when intermingling with each other during the exhibition-annex-reception, many of them came to congratulate me. Now that was a real ego-booster. I started off my talk with stating I was only a talking puppet presenting the labor of love of my 30'ish fellow Cocoon committers, and it really felt that way: being supported by the shoulders of a friendly giant.
During that reception, Cliff Schmidt from BEA (the standards and community guy over there, only recently however after leaving MS where he was doing the same thing) came to talk with me about how they should 'cope' with open source, how they should move toward releasing stuff and building up a community, and how Apache is working along with Sun and IBM in this respect. He was genuinely engaged in the matter and wanted to start a real effort in convincing BEA's management and lawyers to get involved in all this, and I honestly wish him lots of success. It would be cool and confronting (for Sun at least) if another Java heavyweight gets a clue about open source and starts contributing. I also referred him to Brian's Collab.net for infrastructure in that matter. Who knows...?
I also met more extensively with Simon St. Laurent, who turned out to be looking much younger that I thought, and again it was another pleasant conversation on OSCONs, ETCONs, O'Reilly in general and some other interesting things. That black ribbon was now really making a difference, although all the people I met were generally cool and down to earth, and made no big fuzz about them being the voices of our current "metainfoglutified web of conscious thought trains".
After the reception, Marc and I rushed toward the Slug and Lettuce to find out how many people actually showed up for the informally organized Cocoon GetTogether, and after some issues we managed to meet Jeremy Quinn and Richard (In Public), and also Chris and Phil. Pier and Andrew unfortunately didn't show up. After Richard searching in vein for some agenda, much Japanese noodles were enjoyed and non-Cocoon related talks started. Once again, it was no big deal if you weren't there - except for the really big deal of meeting fellow Cocoonies and, because of that genuine community spirit we all feel to be present, being able to skip the formalities and get down to earth very fast, touching base as if we already knew each other for a long time. It's great to see this happen time and time again. I've experienced this several times already, with many Cocoonies, and it's really stimulating, refreshing and thought-provoking. Energy for food for thought. I'm proud to be part of this group.
On Wednesday, I slept a bit longer and got to meet 'someone' who wants to organize an XML and Open Source conference somewhere next year, and wanted to talk with me about XML user groups in general. Again, what had been 'scheduled' to be a quick half-an-hour meeting kept us busy for almost the entire morning, to mutual intellectual and social enjoyment. Time was running out, so I quickly popped by a session about RDF, FOAF and RSS for images on the web by a BBC guy, sat together with Dave again during lunch, and also went to meet and greet the boys from Hippo and X-Hive quite extensively.
Leaving XML Europe left me with a brain full of interconnections, with some really cool contacts and plenty of energy to do useful things with all this in the next few months. I skipped most of the sessions in favor of sitting down with people and I cannot say I regret this. And it was great to hear the same remark from Jacek upon leaving. Unfortunately, we were not able to hear his talk on Gregor since we were on the train already, but he promised to send slides, and I reckon we picked his brains enough to dream up the words he will be accompanying the slides with ourselves.