Hi Jürgen, On Sat, 2 Mar 2019 17:01:55 +0100 Dr. Jürgen Sauermann <[email protected]> wrote: [...] > There exist 2 left-overs from ancient APL 1 times: > > a) ⍴⎕AV = 256, and > b) every defined APL function contains only characters in ⎕AV (which > is the essence of ⎕AV). > > The backgound of a) is that a character fits into a byte and this > property is used by several transfer formats for APL data workspaces. > In the old days there was no ⎕UCS and many APL programs used ⎕AV to > programatically generate APL characters ro to write APL characters to > bytes.
Two questions: 1) Do you know of some NARS2000 ↔ openAPL ↔ GNU APL character translator? I only can use GNU APL, because the other ones do not use Unicode. 2) What is the usage of 'other' characters currently included in ⎕AV? Are they (e.g. @, &, ⇄, µ) reserved for eventual use like '$'? > The )OFF proposal sounds reasonable, I will look into it. Thanks! (That is sort of the opposite of --persist in GNU Octave.) > > Regarding an alternative preferences file, please note that a single > preferences file can contain several profiles (selectable via command > line option -p). Since this file is usually short, I find it more > convenient to have one file with several profiles than several files, > In particular when most options are the same in all files and only a > few differ. Ah, thanks! I did not know about the profiles! > =3= Optimization Questions > > That is just of curiosity: how do the 10 efficiency suggestions by > Bergquist (p.34−37) apply to GNU APL? > Could you provide a reference? I never heard of Mr. or Mrs. Bergquist. BERGQUIST, Gary A. APL Advanced Techniques and Utilities. Vernon/Connecticut: Zark Inc. (1987). Available at: http://www.softwarepreservation.org/projects/apl/Books/APLADVANCEDTECHNIQUESANDUTILITIES That book is so interesting that I (wrongly) assumed that it was known by every APL programmer! :-) (It is in PDF-1.6 format that does not work with some viewers, but pdf2ps + ps2pdf may generate a more compatible file.) > > > =4= Boxing > And when it comes to aesthetics, I find the IBM style with → and ↓ > more APL-like than the NARS style adapted in ]BOXING 21. Yes, I have read that IBM APL 2 also shows the difference between numbers and strings (as ]boxing 29 does, in another way). Thanks, Hudson
