Editorial

by Duncan Pearson

APL96 at Lancaster was seen by many of those that attended as one of the better and more useful annual APL conferences. From the people to whom I spoke I derived a very consistent idea of why this was: APL96 was geared to learning. The emphasis of the conference was on enabling the attendee to learn new skills and techniques that would seriously enhance his ability to do his job. Furthermore the things that were being taught – and I do mean taught – were about the integration of APL into the increasingly complex web of modern computing. Many papers at earlier conferences were about the use of an APL approach to solve certain computational problems. As the various conference reports in Vector will show, the emphasis is now on making those solutions available to other users, on interfacing with TCP/IP, ODBC, EDI &c.

Nowhere was this more apparent than in the brilliant way that Dyadic systems demonstrated for creating OLE servers from APL workspaces. An OLE server is a “black box” that is generally designed to do a specific thing. It provides “methods” (functions) to be run and “properties” (variables) to be queried or set by any Windows tool that can be an OLE client. The inner workings of the server are invisible and immaterial to the user and the data interface is standard, and remarkably like a (possibly nested) APL array. It is the ideal vehicle for the APL developers’ skill, which is with the problems of analysis. The article by Peter Donnelly in this issue explains the process of generating an OLE server from a Dyalog APL/W workspace and should be read by all APL developers who are anywhere near the Windows marketplace.

APL96 was not however the only significant conference this year. June saw a very successful inaugural J conference in Toronto. It lasted two days, was very well attended and was a solid indication of the importance of the language. It was clear that for J, unlike APL, the interest was in the use of the language. Two talks which were repeated in amended form in Lancaster typify this. Chris Burke’s tutorial and Donald McIntyre’s talk on forks were about the language pure and simple. The extra richness of the J language has obviously not yet been fully digested by the community.

Finally, and unfortunately too late for the inclusion of a report in this issue, we have had the APL2000 user conference in Orlando. The programme again seems to have a bias toward the teaching of techniques for interfacing with the outside world. We will have a full report in the January Vector.