APL in Commercial Systems Development: a Survey

Transcribed from the printed issue; not yet reviewed. Read it in the PDF of the issue, page 88

We planned to find out what use is being made of APL for commercial software development. Is commercial APL use prospering—or disappearing?

Are we writing new commercial systems in APL? Are we extending existing commercial systems to meet new requirements? Or are we mostly nursing APL systems into their old age until they can be replaced? If there is new work, in which countries and industries is it being done?

Put another way, we want to see the predictable future of APL (and its offspring) in commercial software development. Not what necessarily will happen, but what looks probable.

Questions¶

To learn that, we need to know the following about each system you know about:

Dialect Perhaps we mean ‘language’: we include not only dialects of APL but also A+, J, K and Niall.

Country Where is (most of) the development work currently being done?

Application For what is the software used?

System identification Owner or vendor, and name or number of the system/s. (You might want to tell us about a group of systems without distinguishing between them.) We will publish none of these, but use them only to ensure we don’t count a system twice.

Developers How many people are working on the system’s code?

Users How many people make use of the system, either through the application’s own user interface or as a server?

Conception In what year did development work start?

Age Not the system’s age in years, but the current part of its life cycle. This is characterised by the work being done, for which we have adapted Shakespeare’s "seven ages of man":

Infancy The system is being developed and is not yet supporting commercial activities of its intended users.

Childhood The system is supporting commercial activities of at least some of its users, though there might still be teething problems.

Adolescence Developers continue to add new function, either to complete the original specification, or to meet new requirements.

Maturity From time to time developers work on the system to add requested functions, or to adapt it to changes in the environment.

Old Age No new function is contemplated. Developers work on the system only to fix operating problems, or to adapt it to changes in the technical or regulatory environment.

Life Support As for Old Age, but money is being spent now on replacing the system.

We’ve distinguished Life Support precisely. Organisations have often designated APL a ‘non-strategic technology’, intending to replace their APL systems as soon as convenient; which then often fails to happen. Many APL systems have spent long years wearing such nooses round their necks.

This reflects two things. One is the success of evangelists for one technology or another in eliciting commitment to corporate technology strategies. The other is APL’s astonishing productivity. Business users with awkward needs who have secured a useful system are unenthusiastic about surrendering it when its replacement may be expensive or less satisfactory or both. Many business users have conducted long and successful defences of their APL systems against a corporate technology monoculture.

So we’ll distinguish as moribund only systems where money is being spent now on their replacements.

Results¶

We thought that an informal survey would answer these questions: that we or our personal contacts would know most of what is going on. That has turned out not be so. (Or perhaps we just hope that there is much more to be found out.)

So we’re summarising here what we’ve learned so far, and are asking you for more information.

Brazil¶

A consultant is using APL for statistical analysis and mapping applications.

Canada¶

A developer has told us of four customer systems in active use: one for economics consulting; the other three are packages sold to the finance industry.

An international data vendor has two major APL production systems based in Canada, with thousands of users worldwide. Systems are mature; support is moving to London.

Denmark¶

A software vendor sells a financial planning package built in Dyalog APL, estimates 5-10,000 users. Mature.

Finland¶

A consultant in Finland sells three packaged systems to the power industry there, one written in J; the others J with APL and Excel front ends. Mature.

Italy¶

A software house in Italy employs 15 developers to work on a financial planning package. Adolescence.

Spain¶

A consultant uses J for financial markets analysis.

UK¶

A major pensions company has been using two systems (APL2 and Dyalog APL) for 6-8 years; both systems are in adolescence; each has 3 developers and 50-100 users.

A major food company has a number of mature applications (Dyalog APL) all but one of which is now on life-support. No new development is likely.

USA¶

A major Wall Street investment bank has been using A+ systems in New York and London since 1988. Systems are now on life support with developers fired or reassigned.

A major commercial finance institution has developed an inventory of 20-40 batch and interactive APL systems over the last 13 years, supporting about 500 users. Development has shifted from APL2 (20 systems) to Dyalog APL. New and mature systems.

A consultant uses J to organise and analyse his demographic data.

Notes¶

Professor Ulrich Küsters tells us that academic use of APL has been growing since 1994 at the Catholic University of Eichstätt in Germany, where he and about a dozen other economists and statisticians use it.

Next¶

Is that it? Or is there more going on? Please use the form on the Vector web site at http://www.vector.org.uk to add more entries following this pattern. We’ll publish an update sometime after the Madrid conference.