Re:SAX

["Ralph M. Deal" <deal@xxxxxxxx>]

Hi, Bob, KLUG,

In response to my failing to find the LINUX link on the soliton site, 
Bob wrote:

> http://www.soliton.com is the site. Scroll down a bit, and you'll see
> "For information on SHARP APL for Linux: "
> 
> and a link there: http://www.soliton.com/Linux

Got to read ALL of those pages.
 
> There's a fair amount of text there, and a link to their download site.
> Download two files, a 2.1 Mb .tgz withall the software and a 6.4 Mb .pdf
> with all the documentation (a professionally organized and written manual
> that was originally created in '85, and has been updated greatly since).

Did you say 700 pages in your presentation?

> >    Thanks for sharing your enthusiasm for APL.
> You are quite welcome. Unlike most bigotry, mine is borne from a quarter
> century of success stories and accomplishments (mine and others) with APL
> and APL-based tools. I use several other languages, but APL has helped
> give me an edge in a very competitive marketplace, and the marketplace of
> ideas.

You made that clear in your presentation.

> >It was somewhat enlightening, somewhat confusing, but made me realize
> >that I needed to learn a new alphabet as well as a different method of
> >evaluation of a string (right to left is strange to me) if I am to use
> >APL as a programming language.
> 
> It is difficult to teach as broad a subject as a whole programming
> language in the hour+ we have for presentations at KLUG meetings, and
> doing more than a walk through of a couple of programs can be very hard,
> especailly if the language is umconventional.

Of course.  I'd love to present Standard Meta-Language (SML) but I suspect
there is no interest.  I'd be faced with the same problem.

> The right-to-left-without-precedence choice made by APL's inventor is
> actaully a very carefully thought out choice, and was adopted for a number
> of reasons. I'll agree that it takes some getting used to, but once you
> get used to it, it's hard to go back! Actually, the same reasoning that
> brought the "ENTER" key (and no "equals" key) to HP calculators is what
> we're talking about here. APL and "reverse-Polish" notations share roots.

Interesting.  I just bought a HP49G and am exploring its somewhat strange
conventions.  I'm familiar with RPN of course from earlier HP calculators
and other languages so I should take to APL easily...

> It isn't accurate to consider APL a dead language; there is ample evidence
> that APL is alive and well, and it is largely in the Computer Science com-
> munity that it is considered "dead". This is largely because computer
> scientists do not appear to be interested in a language that provides 
> notational abstractions for many of the problems they feel are important
> for students to learn. 

Good point.  The question here is what is computer science?  An
engineering course (use APL) or a discipline more closely related to
logic, philosophy (as in AI), mathematics and formal methods (use SML).

> The rest of the world considers many of these to be solved problems 
> (sorting is a good example), and want to go from there to address "real
> world" problems, issues that are of practical use to computing profes-
> sionals, and are not "trivial" computational or data manipulation
> programs. When attempting to solve some of the more difficult problems,
> APL provides an outstanding starting point, both as a notation and
> for inital implementation. Sometimes, if the resulting APL software 
> cannot be adequately optimized  (rare, but it happens), it can be used
> as a programming specification for work in other languages, which are less
> suitable for research and algorithm construction, but perhaps more
> attuned to high performance implementation. 

APL as an algorithmic language? 

> All of this started with the claim made during the thread about the best
> initial programming language. I've made the claim that APL is a very good
> choice in this area, and having taught APL to a number of people, mostly
> to get them productive in programming in a rather short (1-2 quarters)
> time, I beleive I have adequate basis to make such a claim. If we can
> get together enough people to make it worthwhile, I could teach a group
> here, based on SAX and other APL implementations availabel for Linux.

If you do, I'll be one of your students, Bob.

	Best,   Ralph