APL93 Postscript: Don't Stop Thinking About Tomorrow

by Adrian Smith, . (adrian@apl385.com)

After APL91, I upset a few people by wondering to myself if it had been a conference or a reunion. I am sorry, but the past of APL should not be of interest to APLers; let us leave that to the language historians and get on with designing the future.

Toronto-93 could hardly have been a greater contrast. We started with McIntyre showing us how the ancient Egyptians got stuck with an inefficient and redundant script because “it was the Language of the Gods” and change was therefore unthinkable. After 3,000 years, it was replaced by Greek. Typically, he left the conclusions hanging in the air, but the parallel between conventional mathematical notation (ugly, inconsistent, un-linearised, impossible to extend, but sacred) and APL (linear, consistent, extensible) was plain to us all.

From the conference satchels (“APL2 co-operates") to the closing plenary, the accidental theme of APL93 turned out to be the fit of APL into modern windows interfaces and networked environments. There was genuinely exciting stuff on view, both in the formal papers, at the Vendor forums, and at the Exhibition. I left APL93 feeling more optimistic about the future of APL as a notation than I have felt for years.

Conference Highlights

• The Exhibition. This was spacious, accessible and well-attended. The technology was impressive too - it was fun to run through a demo at the Techgnosis stand which just happened to include networked access to an SQL database on one of Dyadic’s machines several booths away. The Soliton balloons were a real hit with the children, particularly when Richard’s had to be rescued from the ceiling by an athletic Dane.
• GUI APLs. Someone commented to me “this feels like Dyadic’s coming-out party". But ISI and IBM were in there too! Dyadic’s beer-and-sandwiches Vendor forum ran out of food, and nearly ran out of beer; IBM also showed a very competent GUI APL2 under OS/2, and ISI had to re-run Eric Iverson’s “Window Driver” tutorial due to popular demand.
• McIntyre on Egyptology. I don’t care if this had anything to do with APL, I just like hearing Donald lecture!
• Gerard “I am driven by Symmetry” Langlet. Gerard has to be experienced - as a live show he is brilliant, even at 8.30am. Some of the images he showed were genuinely stunning, and I left the talk convinced that virus DNA must somehow implement ne\ in hardware!
• Jack Rudd on APL in satellite surveillance. In part for the retrospective, but more importantly for his view of a parallel future, and APL’s role in it.
• Richard Levine’s “Birds of a Feather” on APL utilities. The Toronto Toolkit is already a good product, and after an hour and a half’s hard brainstorming, it should be even better.
• The Conference Literature. Everything from the initial flyers to the Proceedings was well-designed and professionally produced. I suspect that to have material of this quality ‘up front’ is worth at least 50 delegates; it certainly influenced me in signing up early.

The organisation was effective without being officious, and the security (if present at all) was far less overt than at Stanford. The site was almost ideal, with the exhibition area and the lecture rooms within seconds of each other (although some signposts would have been quite useful). The banquet was well up to standard - particularly the gardens - and (of course) Toronto fully lived up to its billing as one of the top conference cities in the world. Where else can you ride a streetcar for $1.30 for a journey of several miles to an olympic-standard pool which is free? They even have districts called Scarborough, Malton and York!

Conference Lowlights

• The panel on GUI standards for APL. It was ridiculous not to invite MicroAPL (who started the whole thing 5 years ago on the Mac), and throughout the session, I felt a building sense of frustration in the audience (articulated in the end by Eke van Batenburg) that these damned implementors needed some users to knock their heads together! Maybe some good will come of it, but it left me with an uncomfortable feeling by the end.
• the Soliton and Manugistics vendor forums. I arrived at the very end of the Soliton session, passing by a stream of disgruntled delegates on my way. Manugistics seem to have perfected the art of self-fulfilling market surveys, and convinced themselves that as long as they do not call their new product APL, the world will beat a path to their door. Maybe they should call it Edsel?
• Please can we have water at coffee-breaks, on speakers’ tables etc. (and indeed at the banquet). Toronto is warm in summer, and if you try to take all your fluid on board as coffee, you end up pickled (or at least I do).
• The Audio-visual stuff. In these hi-tech times, it should be quite normal to rock up to a podium with a Compaq portable, and expect to see some means of projecting it! Generally, a projector was found, but often after a good deal of scurrying about; in at least one case a paper was more or less wrecked when the organisation discovered that it required 3 projectors, and only possessed 2. Several papers were also shunted from room to room at short notice, I assume to fit in with projection requirements.
• The picnic. OK, Stanford was a hard act to follow (and so by the sound of it was St Petersburg), but the Algonquin Island site really was a bit of a grot-hole. The buses were cunningly timed to just miss the ferry (one per hour), and the site was hot, dusty and cramped, with no decent shade. The food was fine, but (as a final offering to the von Neumann paradigm?) the processing was strictly sequential and the queues consequently lengthy.
• the framed speaker’s plaques - gimme a teeshirt every time!
• the loss of Richard’s Soliton balloon to a tree mere yards from our hotel.

I also felt that perhaps the program was perhaps a little bit ‘bunched’; all the stuff on user-interfaces and DDE was on the first day, and all the Statistics came together on the third. I assume that this was to tempt ‘day’ delegates who had a particular interest, but for the rest of us it had the effect of reducing the accessibility of the papers from our own interest group. However I do think it was good to run only two streams, with plenty of parallel slots for rerun tutorials and ‘BOF’ sessions. Not a serious quibble really, I’m sure that however carefully streams are arranged, someone has to make hard choices.