Pick of the Week
Vector 3:2, page 77
Transcribed from the printed issue; not yet reviewed. Read it in the PDF of the issue, page 77
Introduction¶
This little rag-bag of odds and ends is my attempt to convey some of the excitement and sheer fun that we all had at APL-86. I shall draw a decent veil over the exploding Real Ale, the midnight Croquet, and the extraordinary level of design skill applied to the paper aeroplanes at the conference dinner. Suffice it to say that the general intellectual standard at APL86 made your average gathering of academics look like the autumn meeting of the Beano Readers Association.
Many of the papers were tough going; at the end you thought to yourself “I had to work pretty hard, but I think I got some of that”; then you realised that everyone else was thinking much the same! It made me very proud to be a member of such company.
Now for some of the lighter moments – first a selection of my favourite quotes (I’m afraid that many of these are on tape, so the perpetrators would do well to think twice before they commence legal proceedings!):
From the opening plenary:
“I’m grateful they didn’t call it APL”
Ken Iverson on APL2.
From ‘APL2 by Example’:
“The trouble with APL2 is that you take 10 minutes writing the code, and then spend the rest of the afternoon telling your colleagues about it”
“CPU time – if you don’t use it now it’s gone for ever”
Alan Graham
From the panel on implementations of new APLs:
“We’re building interfaces to IBM acronyms that I really don’t know what they are”
“We’ve got to be able to drive mice, and all those other little critturs that run around.”
Bob Bernecky.“Brackets are the cockroaches of APL; have some respect for them, because they’ll probably outlive mankind.”
Adin Falkoff (under extreme provocation).
From the first debate (on the effect of enhancements on the spread of APL):
“The ‘old’ APL is the most fantastic, amazing product the computing world hasn’t seen”
“APL2 was a conspiracy to stop the spread of APL(!)”
“We’re not getting new customers from APL2 – just confusing the old ones”
Richard Nabavi.
From the panel on APL thinking:
“Encode is something you use when you give change”
Norman Thomson on APL thinking.“If you disagree with Alan, it’s like standing at the wrong end of a cannon”
Norman Thomson on Alan Graham
Mexican Snacks at Dyalog APL¶
There was almost a surfeit of APL2 at this conference. I don’t know quite why, but APL2 bothers me a lot; the odd thing is that Dyalog APL looks far more attractive, and I can’t for the life of me see why! Having greatly enjoyed Alan Graham’s “APL2 by Example”, I had still come away with no lasting impression other than that of his extreme cleverness and showmanship. What I certainly did not get was a clear vision of where I could really use APL2 to get my job done better.
That basically means using APL to develop and (much more important) maintain and enhance complex screen-based systems over a wide variety of applications. No-one has yet had to maintain a system in APL2; if they are even half as clever as Alan, I think they could be in dead trouble! Which brings me by a round-about way to the point; the only person at the entire conference who significantly helped me along the APL2 road was John Scholes at Dyalog’s unscheduled “Fringe”.
Of course it was marvellous to escape from conference food, if only for an hour or so, but I cannot credit this as the sole explanation. Much more likely is the fact that Dyalog has been developed by a small team (essentially two people) who actually use APL. Contrast this with APL2, which has been brought about through the enforced union of two APL development groups with very different views on the future of the language. John Scholes could talk convincingly about the value of “Compose” because he knew why he wanted it, and because he was the one who wrote it!
He could demonstrate the power of ideas like “Function assignment”:
. . . because he was the one who had wanted it enough to go to the trouble of inventing it! That is the same kind of motivation that Ken Iverson felt when he first sketched out the ideas of APL; I think that is probably why Dyalog APL feels ‘right’, and why I so much enjoyed those early-evening talks.
The Debates¶
We had two ‘formal’ debates at APL-86; the first acceeded to the view that enhancements to APL do nothing to help the spread of the language, and the second threw out the conjecture that APL is not (yet) a suitable language for Artificial Intelligence. For me, these ranked alongside the panel discussions as events where the rank and file of the APL world could make its views known. There is always a danger at such a prestigious international gathering that the platform will be dominated by the same circuit of ‘names’ year after year, and we ordinary folk will never get the chance to say our bit.
The glory of the debating format is that it positively encourages the outrageous, the controversial, the hyperbolic. You can jump up and say something totally ridiculous, and as long as it ‘scores’ off someone you will very likely get away with it. In the first debate, the floor (and the principals) were to some extent feeling their way; by the end of the second it was beginning to warm up nicely. I hope the committee of APL-87 will take note, and somehow find room for this (very British) tradition in their program. Perhaps a ‘Balloon Debate’ on APL enhancements would be a revealing occassion?
Postscript¶
If you can stand the pace, stay awake, hold your own in the bar, and (at times) work damned hard, APL conferences are a lot of fun. This was my first, I doubt if I will make Dallas, but of course it’s always in Europe in the even years, isn’t it? See you all in ’88.