Dyalog Forums

There seem to be some men at Google & co thinking about wheater a language should still be limited to ASCII characters. Don't we have a solution for them?

ASCII crimps program development, coder says
http://www.computerworld.com/s/article/9194021/ASCII_crimps_program_development_coder_says

Sir, Please Step Away from the ASR-33!
http://cacm.acm.org/magazines/2010/11/100618-sir-please-step-away-from-the-asr-33/fulltext

I read the articles and they are interesting. I learnt APL in 1991 while in College as a grad student. Will the old guard tell me as to why is there so much negativity out there vis-a-vis APL? I keep reading disparaging remarks about APL (for the last 19 years now). What did the "old timers" do or not do to cause this? I have never read similar comments about Fortran, Alogol, PL 1, Ada etc. Can we do some objective introspection?

I'm not an "old guard" but a couple of things come immediately to mind (there will be more)...

0 - Once you get past the superficial differences Fortran, Algol and so forth are essentially the same thing (scalars as the native data entity, looping and so forth). Know one and you can transition relatively easily to others. APL has a different underlying nature, so it led to an unwillingness to start over on the part of many established programmers.

1 - Something that aroused a lot of negativity was APL's order of evaluation rule, especially when it was poorly explained. The "establishment" thinks there's something magically wonderful about BODMAS.

2 - At the time machine efficiency was a paramount concern, "real programmers" didn't like interpreters.

3 - Some APL examples seemed to be deliberately terse and dense.

4 - Some APLers of the time were quite obstreperous individuals (of course, this isn't true nowadays).

Somebody will surely be along shortly with more (and corrections to the above) - but I think that deep down APL was perceived as "a threat", and some of us weren't willing to suffer what we saw as foolish gladly.

Dick Bowman wrote:0 - Once you get past the superficial differences Fortran

That was Fortran 77. Fortran 90, already, had constructs like:

C = A+B

where both A and B are arrays... And Fortran 90 is 20 years old... After Fortran 90, 3 more standards have been published: Fortran 95, Fortran 2003 and Fortran 2008. I used Fortran at the university in 1995. Needless to say, the only Fortran that we could use was a very limited dialect of Fortran 77...