December 12, 2003
@ 12:19 PM

Today is the last day of the XML 2003 conference. So far it's been a pleasant experience.


Attendance at the conference was much lower than last year. Considering that last year Microsoft announced Office 2003 at the conference while this year there was no such major event, this is no surprise. I suspect another reason is that XML is no longer new and is now so low down in the stack that a conference dedicated to just XML is no longer that interesting. Of course, this is only my second conference so this level of attendance may be typical from previous years and I may have just witnessed an abnormality last year.

Like last year, the conference seemed targetted mostly at the ex-SGML crowd (or document-centric XML users) although this time there wasn't the significant focus on Semantic Web technologies such as topic maps that I saw last year. I did learn a new buzzword around Semantic Web technologies, Semantic Integration and found out that there are companies selling products that claim to do what until this point I'd assumed was mostly theoretical. I tried to ask one such vendor how they deal with some of the issues with non-trivial transformation such as the pubDate vs. dc:date example from a previous post but he glossed over details but implied that besides using ontologies to map between vocabularies they allowed people to inject code where it was needed. This seems to confirm my suspicions that in the real world you end up either using XSLT or reinventing XSLT to perform transformations between XML vocabularies. 

From looking at the conference schedule, it is interesting to note that some XML technologies got a lot less coverage in the conference  relative to how much discussion they cause in the news or blogosphere. For example, I didn't see any sessions on RSS although there is one by Sam Ruby on Atom scheduled for later this morning. Also there didn't seem to be much about XML Web Service technologies being produced by the major vendors such as IBM, BEA or Microsoft. I can't tell if this is because there was no interest in submitting such sessions or whether the folks who picked the sessions didn't find these technologies interesting. Based on the fact that a number of the folks who had "Reviewer" on their conference badge were from the old school SGML crowd I suspect the latter. There definitely seemed to be disconnect between the technologies covered during the conference and how XML is used in the real world in a number of cases.


I've gotten to chat with a number of people I've exchanged mail with but never met including Tim Bray, Jon Udell, Sean McGrath, Norm Walsh and Betty Harvey. I also got to talk to a couple of folks I met last year like Rick Jellife, Sam Ruby, Simon St. Laurent, Mike Champion  and James Clark. Most of the hanging out occurred at the soiree at Tim and Lauren's. As Tim mentions in his blog post there were a couple of "Wow, you're Dare?" or 'Wow, you're Sean Mcgrath?" through out the evening. The coolest part of that evening was that I got to meet Eve Maler who I was all star struck about meeting since I'd been seeing her name crop up as being one of the Über-XML geeks at Sun Microsystems since I was a programming welp back in college and I'm there gushing "Wow, you're Eve Maler" and she was like "Oh you're Dare? I read your articles, they're pretty good". Sweet. Since Eve worked at Sun I intended to give her some light-hearted flack over a presentation entitled UBL and Object-Oriented XML: Making Type-Aware Systems Work which was spreading the notion that the relying on the "object oriented" features of W3C XML Schema was a good idea then it turned out that she agreed with me. Methinks another W3C XML Schema article on could be spawned from this. Hmmmm.


Friday, December 12, 2003 1:50:08 PM (GMT Standard Time, UTC+00:00)
"There definitely seemed to be disconnect between the technologies covered during the conference and how XML is used in the real world in a number of cases."

That's why it wasn't well-attended, but I'm sure you already figured that out.
Friday, December 12, 2003 1:52:45 PM (GMT Standard Time, UTC+00:00)
Just a couple of comments (I chair the conference); the numbers at the conference aren't lower than last year, it just looks like it (maybe because it's a bigger convention center?) We don't have final numbers, but it does look a little bigger than last year.

We didn't get any other RSS-related submissions which I found a little surprising... maybe there will be more next year.

Not all the reviewers are a) at the conference, or b) wearing a reviewer ribbon; that being said, the number of reviewers of any topic area is not related to the number of talks in that area that were selected for the conference. The number of talks selected for the conference is related to a) the number of talks in that area that were submitted, b) the grades and comments the reviewers gave the talks (independent of how many reviewers there were), and c) what the planning committee thought conference attendees would find interesting. Since each paper was reviewed by a minimum of 3 people who were interested in that topic area (and most by 4-6), the planning committee got a pretty good view as to which talks and topic areas seemed interesting. As an example, we added the area "client applications" after seeing all the submissions which got high grades from reviewers; it wasn't part of the originally suggested topic areas.
Lauren Wood
Saturday, December 13, 2003 12:47:41 AM (GMT Standard Time, UTC+00:00)
I'd just add to what Lauren said that the proposals are anonymous, so there's no obvious way for a reviewer to discriminate against a certain class of submitters for ideological reasons (e.g., "no big companies.").

Also, the "reviewer" ribbon didn't stick to the "presenter" ribbon on my badge. Maybe others had that same problem, so you didn't notice who the reviewers are. As far as I know, Lauren draws on a pretty broad pool of people to review, and routes articles on particular topics to people she knows have some expertise. I saw *lots* of Web services proposals ... and none of the ones I gave bad grades to looked remotely like they came from IBM, BEA, or MS. If anything, big company people would have an easier time getting on the program because they tend to have a clue on how to give decent presentations.

I suspect that the reason that "there didn't seem to be much about XML Web Service technologies being produced by the major vendors" is that they didn't submit proposals. With all respect to Lauren's great conference, there's probably not that much business value there for them. XML is just the "plumbing" in Web services, and not where the interesting stuff happens (although I believe that it does much of the heavy lifting).

Also, I noticed that the Web services tracks were decently attended but not crowded, while some stuff that has just about zero traction in the market (like XForms) had standing room only. I suspect that it's because people can't AVOID hearing about Web services for the last couple of years, and allocate their conference time to the stuff they don't know much about. I'm sure it's not because the reviewers kept out the good stuff, and while you are right that the geeks who run/review the show tend to be SGML veterans, at least half the attendees appeared to be from the real world. They don't know or care about the "documents not data" ideology, or the "thick/smart clients are evil" ideology, they just voted with their feet for what looked interesting.

Anyway, it was good seeing you again Dare, and you shamed me into posting a series of obervations about the conference into my weblog.

Saturday, December 13, 2003 3:44:44 PM (GMT Standard Time, UTC+00:00)
Aw, wish I was a more frequent flyer, sounds like a good show, and I'd very mcuh have liked to have met you.

Can't resist:
"in the real world you end up either using XSLT or reinventing XSLT to perform transformations between XML vocabularies".
Definitely, and like this where you need to process the content/value it's currently the best bet for getting RDF/XML out of arbitrary XML data. But RDF allows mixing of, and powerful relationship modelling between *RDF* vocabularies. Different job.
Wednesday, December 17, 2003 1:55:36 PM (GMT Standard Time, UTC+00:00)
The sheer size of the conference center in Philly definitely made the show feel smaller than it actually was. Also, there was some serious distance between meeting rooms. This lead to small groups converging on coffee/tea during break times in different time zones.
I thought the show was great. I did not get to as many talks as I would have liked to but, as always, it has sparked ideas in my head that will keep me cogitating through the winter months.
Comments are closed.