Author Message

Homer Smith,

I don't know how I can make it any plainer.

APL2/PC for DOS is NOT a frozen product.  In your conversations
with my manager, you must have misunderstood and/or he must have
been unclear.  Please see the article I posted yesterday for
an explanation of our focus in 1993.

David Liebtag
IBM APL Products and Services

Mon, 29 Jan 1996 00:08:29 GMT  

I recieved a message from Curtis Jones regarding the recent thread
on APL2. I think this info is of interest to APL'rs generally so
I asked if I could post. So here goes...


Yes, I think posting the reply would be OK.  Just don't say I'm in APL
development.  (I'm an engineer/scientist in ADSTAR.)  Don't think I
said anything that sounds like a policy statement, but this is one
place where a disclaimer that my own, possibly slightly faulty
recollections shouldn't be taken as company proclamation seems valid.

===================original message

>>     However it should be stated up front that APL2/PC is a frozen
>>product, originally made by a company that is now out of business,
>>Winchester Scientific, Cambridge England.

>  I it's not accurate to say APL2/PC is a frozen product.
>  Within the last year APL2 has been rewritten from the IBM portable
>  proprietary language it was originally coded in (in England) into C.

Think there's a little confusion about nomenclature here.
The APL2 for DOS is called "APL2/PC".  The other, which David Liebtag
described to the APL BUG, is "APL2 for OS/2" or "APL2/2".

The statment about the conversion from the proprietary language to C
does correctly describe the development of APL2 for OS/2.  The
original work on the proprietary language was actually done in Madrid,
Spain.  But a lot of important work was done in Winchester:

APL2/PC really started in the early '70s at IBM's Madrid Scientific
Center (Spain) when Manuel Alfonseca and several colleagues wrote
an APL interpreter for the IBM System/7.  You've probably never
heard of a System/7.  It was designed to control machinery -
lab testers, traffic lights, grocery store refrigeration systems...
Manuel, I think, had the idea of writing the interpreter in what
he called "Intermediate Language" or IL, and then writing a small
translator to generate machine code for the specific machine from
the IL source.  So when the IBM Series 1 came along, it didn't take
long to get APL running on it.  Then the IBM PC arrived, and
Manuel's group used the IL source to put an APL interpreter on it.
That was sold as APL/PC Version 1.  I think it was the first
product for the PC to use the math coprocessor.  It also had a
number of bugs, arrays were limited to 32 KB (bytes, not elements!)...

David Selby and Ron Wilks at IBM's Winchester Scientific Centre went
to work on improvements.  I think Dave more or less organized the work
and wrote a bunch of APs.  Also incorporated work done in other parts
of IBM.  Manuel made substantial improvements to the interpreter, and
Ron documented it all and put together a package that could be sold as
APL/PC Version 2. Unfortunately, it never got sold in this country.
:-( I bought a copy myself from an outfit in Boston that imported
APL/PC Version 2.

When APL/PC Version 2 was out, they got APL2 going in the
IL interpreter.  This led to the current APL2/PC product.  The IL
source was also used for the APL2/6000 that was developed in
Kingston, New York.

In these APLs for DOS, the source for the APL interpreter was
written in IL and the IL was translated to machine code for the
target machine on, I think, a mainframe.  A few parts of the
interpreter may be written in assembler for better performance.

APL2 for OS/2 is a new project being done by the APL development
group at the Santa Teresa lab.  I think they have used the IL
source a lot for their C code for the interpreter.

  . . .

>  over having had nothing at all. My understanding is that at Palo Alto
>  3 people are working full time developing more stuff. I have no
>  idea if these 3 are among the many waiting for the axe :-(, but

It's the Santa Teresa lab, and the APL developers are highly
unlikely to get the axe.

>  it's been rumored that IBM is still keen on APL2. You can expect
>  that AGSS will be marketed more aggressively than ever (which is
>  not saying alot) once the promotion ends.

True, I think.

>  -emmett

I trust you've picked up a copy of the IBM Systems Journal,
Vol. 30, No. 4 (1991) with all the articles about APL.
You might enjoy Alfonseca, Selby and Wilks "The APL IL Interpreter
Generator" starting on page 490.    Curtis

Thu, 01 Feb 1996 12:20:20 GMT  
 [ 2 post ] 

 Relevant Pages 

1. APL2/PC display font problem

2. DIF files & APL2/PC

3. APL2/PC display font problem

4. COMPLEX.QQ : necessary adjustments for use by APL2/PC

5. 32-bit APL2/PC under Windows

6. APL2/PC; ATF-file format

7. APL2/PC and DRDOS

8. memory managers & APL2/PC

9. APL2/PC Exchange Assembly files

10. 32-bit APL2/PC

11. IBM apl2/PC 1.02

12. APL2/PC and subdirectory-library mapping


Powered by phpBB® Forum Software