My first day at the Python conference, but the second day of the conference. Greg and I thought everything started at 9PM, but the first papers actually started at 8:30, so we arrived a bit late. Greg went off while I sat down to do a last check through my MX slides and to finish up my RELAX NG slides.
My talk went reasonably well, though there weren't very many questions and I kept tripping almost every time I walked over to the projection screen to point something out.
My favorite memory from this conference, though, will likely be from before my talk, when a man came up to me as I was standing around and said "Hello, I'm Tim Berners-Lee. I won't be able to attend your talk, but I just wanted to say thank you because we use some of the software you've written, such as amkCrypto." A truly great moment for me. (This makes me feel bad for not releasing a new version of amkCrypto in so long.)
Kevin Jacobs, whose talk "Integrating Business Information with Python" came after mine, was far more interesting than I was, and his presentation gave me some interesting ideas for our project. Jacobs's group develops a Web-based reporting and tracking system that's a front end to large corporate systems such as SAP R3, and users do similar things to ours: getting reports, entering changes, that sort of thing. They have stricter browser constraints for their users, requiring Netscape 6 or IE 5.5, meaning that they can rely on using dynamic HTML more than we currently do. This has some advantages, as they can present report tables that are sorted using client-side DOM, and use fancy DOM-based widgets such as a browser-based spreadsheet. Maybe we can use similar widgets for the MEMS engineers, and just be sure they're using current versions of Netscape.
Lunch was unremarkable, as these catered hotel lunches almost always are, and this vegetarian meal was especially aggressive in its blandness. Afterwards, I went to the web services panel, which turned out to be pretty interesting, starting out with web services but ranging afield to other topics such as RDF. I wanted to leave early to go to Martin von Löwis's talk on choosing a standard parser for Python, but thought it started at 3PM when it actually started at 2:30; 3PM was when the session ended. Oh well; at least I have the paper in the proceedings.
Berners-Lee's keynote speech, "Webizing Python", came next and was astonishingly disappointing. The slides were obviously written in a hurry and not proofread, as they were filled with typos (if you go look at them, note that "etgas" is a typo for "e-tags"). It started out by suggesting the addition of a RDF-style labeled graph data type to Python, so you could write graph = {sky color blue, grey}. Another suggestion was identifying Python modules with URIs, but I think the audience missed the fine distinction that this suggestion is not the same as suggesting that modules are actually directly downloaded from those URIs, as shown by someone's question about the security aspects of this. General reaction was not very favorable; for example, Greg acerbically noted that the example of "sky color blue" was an appropriate one.
The primary fault of the talk, though, was that it proposed a bunch of changes and additions -- a solution -- without ever explaining the problem. Someone later suggested that a brief history of the Web and a discussion of its current problems would have made the need for, and purpose of, the Semantic Web clearer and motivated the resulting language suggestions. (TBL's hurried speaking style didn't help -- I got the impression, listening to him during this talk and during the web services panel, that ideas crowd upon him faster than he can clearly express them.)
Afterward, the prize for best paper was given to the ViPEr paper by Michel Sanner and various co-authors, and then the Frank Willison Award for Contributions to the Python Community was handed out. To me, in fact.
The actual presentation of the award narrowly avoided disaster. After tripping all through my talk in the morning, I put a seal on the day by almost tripping in a spectacular way when going up to get the award. The speaker in this room was on a little podium with a stepladder leading up to it. A rather frail, bendy stepladder. Now, I don't so much walk up stairs as bounce up them, so picture it: I come up the side aisle, stride up to the podium, and bounce from the ground onto the second, top, step which buckles under me. I thank my lucky star that I didn't pitch forward and collide with the O'Reilly editor presenting the award, but instead managed to actually remain stable. (For much of the rest of the conference random people would stop by and congratulate me. It was quite a nice feeling.)
The rest of the evening proved kind of confusing, and I wound up not getting any dinner at all. After going out to the car to leave the framed certificate for the award (picture forthcoming in a few days), I came back and found that Greg and Anton had scampered off for dinner to parts unknown, though I didn't realize this at the time. There was a ZODB4 BoF session that we wanted to attend, so, while I followed the Pythonlabs and O'Reilly groups into the hotel's restaurant, I dashed off after 10 minutes without ordering anything.
The BoF session proved disappointing. It was intended to be a discussion of what a general persistence API for Python should look like, so that object-relational tools such as MiddleKit and object tools such as the ZODB could be used more interchangably, but as almost all of the attendees were potential users, not potential implementors, it turned into simply a summary of the existing ZODB interfaces with some suggestions about the changes needed to make them more general.
As the BoF was ending, Greg and Anton reappeared. Anton hung around a bit and then went home, while Greg went off to the Python Software Foundation meeting. I decided not to catch a ride home with Anton, as I still had the car keys (from taking my things out to the car, remember?), but didn't know where the PSF meeting actually was so I could find Greg. Instead I hung around and had a nice chat with Andrew Dalke while getting steadily hungrier and hungrier.
Two hours later, I could stand it no longer and went off in search of the hotel's vending machines, but ran into Greg and everyone else, just emerging from the PSF meeting. It was apparently monstrously dull; a new board was voted in, there were lengthy discussions of the PSF's intellectual property agreement, 11 new people were to be invited to join, and the conference was discussed (or maybe "flamed about" would be more accurate). Apparently I'm one of the 11 people to be invited, but I turned down the invitation last time and, after hearing how agonizingly dull the meeting was, I'm going to turn it down again.
And so we went home, where I finally got something to eat! (Oatmeal.)