rob@conexch.UUCP (06/03/88)
For those interested, IBM announced (privately) today APL2 for the PC. APL2 is a full blown implementation of IBM/APL2 V2. Not Version 3 as the mainframe recently was released. I have seen a demo at our local SIGAPL chapter by James Wheeler of IBM. It's about 140K + all the auxiliary processors of your choice. It even looked rather fast. I can't wait to get it onto my '386/20! Price is said to be $500.00. Availability? WHo knows. Since James Wheeler called my boss (Roy Sykes) to tell him the news, I suppose Jim was simply telling Roy IBM had made the decision to market the product. IBM has been using it internally for quite some time though. I hope this news is good news to all of you who want nested arrays, user-defined operators, et all for the PC. -- "Worship the Lord your God, and serve him only." Mat. 4:10 Robert Collins UUCP: ucbvax!ucivax!icnvax!conexch!rob HOMENET: (714) 995-7344 UUCP: uunet!ccicpg!turnkey!conexch!rob WORKNET: (714) 229-0284
amull@Morgan.COM (Andrew P. Mullhaupt) (11/08/89)
There have been a couple of requests for this information from far-flung corners of the globe. (New Zealand is flung pretty far from NYC, for example.) I have bounced the mail a couple of times unsucessfully, so here's the post: APL2/PC is a product of IBM. (International Business Machines Corp.). It has a version that runs on 386 (32-bit) machines. The language is very close to APL2, which runs on 370 style mainframes, with one or two minor exceptions. I use both APL2/PC and STSC APL*Plus. They work OK, but like almost all APL implementations, don't expect speed. APL2/PC has a bit more zip than the version of APL*Plus I have, but both pale before the strength of any good compiler. Quad NA is not available for APL2/PC. Another alternative for interfacing with compiled code is provided, but I have never got it to go at all. This is probably because I don't run PC-DOS, and use the IBM Professional FORTRAN products. I much prefer Microsoft FORTRAN 5.0, and would rather use Pascal anyways. (Are you listening, IBM?) APL2/PC is useful to me predominantly because I program extensively in APL2 on a vector 3090 all day long, and it's nice to have the small version of the language running on my laptop for algorithm development. APL2/PC helps keep me from using my home terminal too much, because the PC I have is ergonomically about three generations beyond the 3279 terminal. APL2/PC doesn't run in a DesqView 386 window. Qemm gets in the way. there is also no OS/2 version yet. Perhaps someday. Later, Andrew Mullhaupt Disclaimer: Any opinions expressed above are not necessarily those of Morgan Stanley and Co, Inc.
liebtag@stlvm20.vnet.ibm.com ("David Liebtag") (05/14/91)
From: felscher@MAPLOG.INFORMATIK.UNI-TUEBINGEN.DE
Walter,
I asked one of the APL2PC developers your questions and got
these answers (slightly reworded.)
1. Having installed 386MAX on a PS2/80, invocation of
both the 16bit and the 32bit version of APL2/PC
caused the machine to stall. Apparently, certain high
memory areas should be kept free. Is it known
which ones, or which other remedy might be taken ?
386Max will run with the current 16 bit IBM Internal Use Only level
(which will be available to customers in the fairly near future)
but will not run with the currently available product level.
The 32 bit level of APL2PC will never co-exist with 386Max
because it uses the VCPI interface the same as QEMM. Supporting
co-existence would require a major rewrite of the interpreter
control code.
The 16 bit level will require a 387 to operate as I {the developer}
suspect the emulator won't function. It {the emulator} switches a
bit in the CR0 register to make the CPU inform us {the interpreter}
every time it sees an FPA instruction. But, you must be running
priviledged. In normal DOS mode you are, but when one of the
386 styl programs runs, you aren't in control anymore.
We can provide more detail if required.
2. Having installed the cache programm HyperDisk 4.21, the
16bit version worked allright, but the 32bit version came
up with immediate WS FULL and no response to commands -
though, fortunately, )OFF still did work. Any known
remedy ?
This sounds like a case of Hyperdisk not using a recognized
method of identifying the highmemory area it has allocated.
We will work with two methods: Remap of int 15h or Vdisk.
The latter is used by such programs as IBMCACHE.
If you need further details, please write another article and
I will try to get more information.
Regards,
David Liebtag
IBM APL Products