#I Can Succeed with APL - and So Can You
#J: Life After The Workspace
#The Epistemology of APL2
#Zero - A Practical Programming Language
#APL and the Media (control)
The tenth in the irregular series of APL as a Tool of Thought made its welcome comeback after more than three year absence; sensibly located, sensibly priced and sensibly located as usual. The advertised programme looked good and enquiries of Cheapskate Travel revealed that airfares to picturesque Newark were at an all-time low (air travel taxes were, of course, at an all-time high - why can't the corporate expense accounts in Business, let alone First, be charged on a percentage of their ticket price rather than the tax be a flat amount for all?).
Tool Of Thought comes without the baggage of the bigger conferences; no puffery, no banquets and no Norwegian Folk Dance Cultural Events. What you get is what's on the packet - a day of APL discussion and presentation.
For the meat of the day, four parallel streams were presented; two of the sessions were repeated - but this still meant that there were 18 sessions from which four had to be selected. Ten of the presentations were paralleled by printed material in the Proceedings and although a follow-up CD is being proposed I think this proportion reflects poorly on both organisers and presenters (before rushing in to attack the easy role of the critic, please bear in mind that I have more experience of both organising and presenting than some would like to credit). I believe that presenters need to take these events, and the opportunity to document their thoughts, much more seriously than they have at times; also that organisers need to bring more pressure to bear on some of the more performance-based presenters. Of course, these remarks apply more widely than just to this single event.
(Material in Proceedings)
The Eric Baelen of 1998 is much more upbeat than the Eric Baelen I remember from earlier in the decade. Eric's message is a straightforward one; APL has been responsible for a great deal of success in the past and it can be a vehicle for much more success in the future.
He points out that among the advantages of APL (and, on his behalf, I will include all APL-related variants - as always) are:
Eric also points out, quite relevantly, that what counts (at least in some parts of the world) is Business success; and that if we want APL to be successful in this arena we will have to concentrate on business issues rather than technical issues. Above all , the importance of spreading a success-related marketing message.
(No material in Proceedings)
One of J's major beaks with the APL tradition has been the abandonment of the workspace and the arrival of scripts as the primary vehicle for defining long-lasting objects. My own casual work with J has helped me understand that this is different, but not necessarily greatly better or worse than traditional APL. I attended this presentation in the hopes that J-guru Burke would offer a more gripping case for scripts and some insight into how we might usefully harness their more useful features.
Although one or two things became clearer, the talk got rather bogged down on detail and failed to progress as far as I (at least) would have liked. Although I see some gains from scripts I still wonder whether we might be seeing a situation in which the J-ists have over-leaned their development environment and are beginning to become defensive about their past decisions. What I would like to see, at some point, is a hard-hitting presentation which clearly shows that J scripts have have attractive solutions to some of the issues which serious application development requires. APL, as a whole, has never offered a language-level application development environment at a serious corporate level and I think we all need to be more aware than we are of how alternative languages like VeeBee, Delphi and Java are presented to the practicing programmer (it's all very well to scoff, but not when the opposition has more to offer than you have). For now, I still think that John Baker's approach in which J verbs are held as explicit entries in a database has more to commend it for serious development than bare J scripting; I wanted Chris's talk to convince me otherwise - it didn't.
(Abstract only in Proceedings)
One of the little-appreciated strong points about APL2 is that it has a fairly firm theoretic underpinning (J proponents might suggest that they have a stronger basis, but that is - for now - beside the point), and this is a strength when compared to some of the kitchen-sink appendages which have been found in some APL implementations over the years. Phil Benkard has long been one of the prime theoreticians behind APL2 - even if not directly involved with the implementation group. One of the pleasures of APL events over the years has been Phil's ability to help a half-congealed understanding turn into proper comprehension.
More people should listen to Phil and read his papers; especially some of the misinformed dimwits who call themselves APL programmers on the basis of being conscripted out of the Cobol development warehouses of the early eighties. The simple truths told by people like Phil and Ken Iverson have been ignored for too long.
Phil's story this time was of how notational and conceptual advances in APL had turned 'difficult' problems into simple ones. More specifically he explored investigations which he had carried out into rediscovering the underlying concepts of object oriented programming by developing some relatively simple APL tools, noting along the way how concepts such as multiple inheritance would just fall naturally into place.
It was a great pity that Phil's material didn't find its way into the Proceedings.
(Material in Proceedings)
It seems that every once in a while someone needs to build 'the new APL'. Of course, we know about J, and proprietary languages such as A+ and K are used by those who have access to those particular doors. Alan Graham's 0 is the newest kid on the block (unfortunately still so new as to be still just around the corner) but differs from A+ and K in that it is intended to be available for all.
Alan has made a very persuasive case for 0, and the examples illustrated (the presentation followed the written material fairly closely) that many of the impositions of APL are neatly removed.
We didn't get a demonstration though, and it's still unclear precisely how 0 is to be positioned in the range of programming languages which we can choose from when solving a particular problem. Questions which are still unanswered revolve around how 0 might fit into a world which demands GUI applications and how it links into files and databases. My personal opinion is that APL (and derivatives) might best concentrate on being accessible calculating engines than have to carry round all the baggage of a 'total solution'; 0 might just fit this bill.
0 has been trailed for some time; I want one -soon, please.
(Material in Proceedings)
I don;t know if I was punch-drunk, or what, but this presentation sailed cheerfully in one ear and out of the other. An upbeat proposal that APL could and should be used for developments based around SMIL (Synchronized Multimedia Integration Language) through the vehicle of the Microsoft Media Control Interface, all in the service of creating Distance Learning applications. Some demonstrations of how it all fitted together would have been good, and without them the presentation seemed like not a lot more than cheerleading. Perhaps it was all so simple that showing us would have given away all the trade secrets.
Nevertheless, the talk clearly brought out one of Eric Baelen's suggestions for APL success; find a market niche, exploit APL's rapid development and make the niche your own before 'conventional tools' could deliver.
(Material in Proceedings)
Chestnut time, yet again. APL clearly is a valuable tool in some parts of the insurance industry; indeed we could make the case that APL is an essential tool for this industry and one which is making a comeback as the limitations of spreadsheet and package technology become more apparent.
But the old 'APL is 100 times more productive' argument? Let's give it a rest. Application development is more than programming. Let's also become more aware of what the development environment looks like for some of those 'awful' languages. Java, for one, is clearly oversold (corporate programmers do what corporate managers tell them, and corporate mangers buy what hucksters want to sell them). But take a look at the development environments for Java; compare these with what you have for APL - look at them both with the clear vision of the novice; which one looks like you are going to build reliable and effective business solutions?
Gary left us with some remarks about a GUI application development tool which he is building and which he would like to use as a vehicle for making APL irresistible to the world at large. I want one of these, as well. Maybe it will fit with 0.
Truthfully, despite some of what you read above, this was a very upbeat event; it reflects an optimism about APL which has begun to be apparent not only in events like this but more widely in the past few years. We don't have to be defensive about APL in the way some of us have been; there is clear evidence of APL (with no sales effort) displacing some highly touted and expensively marketed corporate wares.
Events like Tool Of Thought are good - we shouldn't be parochial about them. Several of us made the trip from Europe - more should have.
By the way, making my way back to Newark as the PATH train wove its way through the New Jersey freight yards (Dogon Research spares no expense to give its employees a luxury world travel experience) I saw an APL Stack train. Don't know how we nominate somebody for an Outstanding Achievement Award any more, but I think the guy who wrote APL on all of those container trucks deserves some sort of recognition.
This page last updated 14 March 2013 (repaired links)
© Dick Bowman 1997-2013
Return to the J\APL Contents Page