Technical Correspondence (Henriod, Vesperoni, Sullivan)

Transcribed from the printed issue; not yet reviewed. Read it in the PDF of the issue, page 103

This column features letters which are unlikely to interest readers who do not already know APL. Writers are requested to observe the requirements for the inclusion of APL code in VECTOR. All correspondence should be addressed to the Editor, who will forward technical letters to the technical editorial team.

From Claude Henriod, 19th December 1985¶

Sir: The article “Steps toward a better APL” in VECTOR 2.2 page 91, has especially retain my attention. Each programming language has its own characteristics, beauty and lack compared with the other languages.

Why disfigure the APL with a simulation of Basic?

In short:

the APL presented code contain some missusage
some functions are not in a good APL (NEXT and THEN)
the simulator is a bad Basic

Consequent, the two languages are diminished.

Please maintain APL as it is!!

I have not sufficient vocabulary to give my disappointment in English, accordingly you will find an annex with more technical details in French, and APL.

I suggest to you to transmit the document at Ivor Kenson.

Claude Henriod
Manteiller CH – 1965 SAVIESE

Ed: It seems that some of Ivor Kenson’s Anglo-Saxon humour has failed to come across in translation. Claude’s technical details appear later in this issue.

From: Eduardo Vesperoni, 7th January 1986¶

Sir: As an affectionate reader of VECTOR, I am glad of submit an entry for your competition on range union, just for gratify you (I hope) with an entry also from Italy.

Yours,

Eduardo Vesperoni
Vefreria Parmense Bormioli Rocco s.p.a.
Casella Postale n.354 Parma

From J Sullivan, 11th December 1985¶

Sir: I enclose my competition entry, which was written in IBM APL2 release 1.0 and run on an IBM 3084. However, the functions also works on our IBM PC running APL*Plus Release 4.0.

This function does not work if the shape of the input matrix is 0 x 2; however the competition specifies that input will always be a 2-column integer matrix (which implies it is never null) so special checks will not be necessary (so I haven’t included any for the null-input condition).

You might like to know that about a week after developing this function I suddenly found a use for it; so your competitions do not only while away the time but are useful as well.

Yours faithfully,

J Sullivan
Research Officer
Statistics & Market Research Section
National Westminster Bank PLC
41 Lothbury
London EC2P 2BP

Ed: Thanks for pointing out the usefulness of the competitions! We think that an integer matrix includes a 0 row matrix. It also says much for the portability of APL that you can run your code not only on such different sized machines but using versions of APL from different suppliers.