mercer@mozart.uucp (Randall Mercer) (11/18/88)
Is APL a dying language?
rjfrey@kepler1.UUCP (Robert J Frey) (11/19/88)
In article <723@convex.UUCP> mercer@mozart.UUCP (Randall Mercer) writes: > >Is APL a dying language? Not for me it's not! I was first exposed to APL on an old IBM 5100 "portable" computer over a decade ago. Since then it has remained my language of choice. I'm a mathematician by training and find APL the most natural tool for express- ing and implementing my ideas. Prototyping is extremely rapid, and once one gets used to it APL is quite "readable" and maintainable. Modern versions such as IBM's APL2 or Dyalog APL which implement nested arrays, user-defined operators and similar extensions are extremely powerful languages capable of both numerical and symbolic processing of great generality. Within IBM many people are using APL2 in their AI work. I've done a fair amount of Lisp programming myself and would be hard-pressed to find a example of some- thing in Lisp that would be more difficult to implement in APL2. Right now I'm using Dyalog APL. It's optimized for a UNIX environment, and, unlike many APL's, it is easy to link in external routines and escape to the operating system when you need to. ============================================================================== |Dr. Robert J. Frey | {icus, spl1, dasys1}!acsm!kepler1!rjfrey | |Kepler Financial Management, Ltd.|------------------------------------------| |100 North Country Rd., Bldg. B | The views expressed are wholly my own and| |Setauket, NY 11766 | and do not reflect those of the Indepen- | |(516) 689-6300 x.16 | dent Republic of Latvia. | ==============================================================================
max@claris.com (Max Rochlin) (11/20/88)
From article <723@convex.UUCP>, by mercer@mozart.uucp (Randall Mercer): > > Is APL a dying language? In summary, NO. Big APL users include Insurance and Casuality companies, Banking, and Astronomers. I may be biased, though, as I am a big fan of APL. ------------------------------------------------------------------------- Any statements made above are my personal opinion and may not shared by my current or past employers. ( feel free to zero rho my response ) /---------------------------------------------------------------------\ |UUCP: {ames,apple,portal,sun,voder}!claris!max Applelink: Rochlin1 | | {ames,apple,portal,sun,voder}!claris!madmax!max [home system] | |Internet: claris!max@ames.arc.nasa.gov Phone: 415-960-4052 \---------------------------------------------------------------------/ .
dag@fciva.FRANKLIN.COM (Daniel A. Graifer) (11/23/88)
In article <723@convex.UUCP> mercer@mozart.UUCP (Randall Mercer) writes: > >Is APL a dying language? I doubt it. I have a quandry in my work. I'm the only one here who knows APL, so anything that can possibly be done in PC-type modelling environments gets done there. The really hard stuff gets done in APL. I've heard that the broadcast news services and pollsters were all using APL based software to analyze election and poll results this year. Sounds like an ideal application to me. -- Daniel A. Graifer Franklin Capital Investments uunet!fciva!dag 7900 Westpark Drive, Suite A130 (703)821-3244 McLean, VA 22102
pool@blake.acs.washington.edu (Jonathan Pool) (11/25/88)
APL may not be dying, but STSC has refused to commit itself to upgrading its APL*PLUS for the Macintosh so it will work on a Mac II. If you read the usual Mac user magazines you will almost never see a mention of any APL for the Mac.
bandy@well.UUCP (Andrew Scott Beals) (11/26/88)
[please pardon my flames as I do not hold the mac to be a sacred cow] [oh, and btw, if anyone could recommend a good apl for the mac...] In article <264@blake.acs.washington.edu> pool@blake.UUCP (Jonathan Pool) writes: >APL may not be dying, but STSC has refused to commit itself to >upgrading its APL*PLUS for the Macintosh so it will work on a Mac II. Could it be that some other company has stolen their mac programming talent? In my last job search, I found a number of shops that were proud of raiding "Entire departments" from other companies. In short, if your program doesn't work on a Mac2 or under Multifinder, you're breaking the rules (such as writing directly into the screen). Programs that do this aren't "too hard" to do upgrades. [STSC: I do take contracts.] >If you read the usual Mac user magazines you will almost never see >a mention of any APL for the Mac. Of course not, it doesn't have anything to do with Hypercard, Microsoft Word, Bill Atkinson or Jean-Louis Gasse'e. The mac magazines don't give a hoot about Real Programming or even Real Programming Tools. The only decent programming resource is MacTutor. -- for those of you who don't trust the headers: bandy@lll-crg.llnl.gov or {pacbell,lll-winken,hoptoad,hplabs,apple}!well!bandy
max@claris.com (Max Rochlin) (11/26/88)
From article <264@blake.acs.washington.edu>, by pool@blake.acs.washington.edu (Jonathan Pool): > > APL may not be dying, but STSC has refused to commit itself to > upgrading its APL*PLUS for the Macintosh so it will work on a Mac II. > If you read the usual Mac user magazines you will almost never see > a mention of any APL for the Mac. There are two reasons that STSC has refused to commit to upgrading APL*PLUS for the Macintosh. The first reason is that they bought another Mac APL added some of the APL*PLUS features and called it APL*PLUS. I don't know if they have the source code to update. The second and more important reason is that STSC was bought out a while ago and not very interested in the APL end-user community. If you want to time-share and buy services from them they are very interested in you. STSC isn't the same company it was years ago ( but then again, neither is most any company in the computer industry). Al Rose left long ago. I just hope that IP SHARPE comes out with a Macintosh APL. (All the opinions expressed above are mine and do not represent my current employer, my past employer, or any future employers.) Max Rochlin /---------------------------------------------------------------------\ |UUCP: {ames,apple,portal,sun,voder}!claris!max Applelink: Rochlin1 | {ames,apple,portal,sun,voder}!claris!madmax!max [home system] |Internet: claris!max@ames.arc.nasa.gov Phone: 415-960-4052 \---------------------------------------------------------------------/
metzger@mozart.uucp (Bob Metzger) (11/26/88)
perhaps a more interesting question: Is APL a growing language? A good metric would be: How many people do you know who have STARTED using APL in the past 5 years? I suspect that current users STARTED using APL during its heyday (as measured by attendance at APL conferences) -- 1976-1983. /Bob
rjfrey@kepler1.UUCP (Robert J Frey) (11/26/88)
In <264@blake.acs.washington.edu> pool@blake.UUCP (Jonathan Pool) writes: > >APL may not be dying, but STSC has refused to commit itself to >upgrading its APL*PLUS for the Macintosh... I think that's unfortunate, but APL is STSC's main product, so I wouldn't assume that this is a reflection of their assessment of the APL market in general. Right now STSC is heavily involved in its APL products for MS-DOS and its UNIX-based APL. They may not want to spread themselves too thinly. Any comments from someone at STSC? ============================================================================== |Dr. Robert J. Frey | {icus, spl1, dasys1}!acsm!kepler1!rjfrey | |Kepler Financial Management, Ltd.|------------------------------------------| |100 North Country Rd., Bldg. B | The views expressed are wholly my own and| |Setauket, NY 11766 | and do not reflect those of the Indepen- | |(516) 689-6300 x.16 | dent Republic of Latvia. | ==============================================================================
rjfrey@kepler1.UUCP (Robert J Frey) (11/26/88)
In article <7082@claris.com> max@claris.com (Max Rochlin) writes: >..STSC was bought out a while ago and not very interested in the APL >end-user community...STSC isn't the same company it was years ago... > >I just hope that IP SHARPE comes out with a Macintosh APL. We have been using a product called Dyalog APL developed by Dyadic Systems, Ltd., a U.K. outfit. It is an excellent implementation and, unlike alot of other APL's, is well-integrated into its environment. I've used IP Sharpe's and STSC's APLs for UNIX and Dyalog beats them hands down! I've also used Sharpe, STSC and IBM's APL2 in an IBM MVS environment. One nice feature of APL2 is its quadNA facility which permits you to easily and transparently link in external functions into your workspace. Although Sharpe and STSC also have some provision for this it is neither easy nor transparent. Dyalog APL has a quadSH facility which closely mirrors APL2's quadNA (probably the implementation is quite different). This means you mix APL and C in your work with very little pain, and it gives you a natural path from proto-typing to implementation. Alot of applications start out all APL and then have the appropriate functions re-implemented in C as they mature. ============================================================================== |Dr. Robert J. Frey | {icus, spl1, dasys1}!acsm!kepler1!rjfrey | |Kepler Financial Management, Ltd.|------------------------------------------| |100 North Country Rd., Bldg. B | The views expressed are wholly my own and| |Setauket, NY 11766 | and do not reflect those of the Indepen- | |(516) 689-6300 x.16 | dent Republic of Latvia. | ==============================================================================
prins@prins.cs.unc.edu (Jan Prins) (11/30/88)
from article <7082@claris.com>: > ... STSC was bought out a while ago and not very interested in the > APL end-user community. If you want to time-share and buy services from them > they are very interested in you. > Max Rochlin (claris!max@ames.arc.nasa.gov) Correct me if I'm wrong (or even if I'm not), but I thought the sequence of events was that (1) STSC was acquired by CONTEL essentially as a well- connected (network-wise) data processing center, and (2) was re-acquired from CONTEL by STSC employees when diversification into the computer industry became less attractive a few years back. STSC still sells time-shared APL (and other service) cycles, but also supplies APL*PLUS systems for various configurations of end-user machines. I imagine they sell where the money is: APL*PLUS/PC for PCs and (a smaller market) APL*PLUS/UNX for various unix systems and workstations. The latter is a full implementation of APL with nested arrays, native and component files and plenty of unix interfacing support (Quad-SH, etc.). This last summer STSC released APL*PLUS II which is a hybrid of the two products listed above for 80386 machines: a fully modern APL system with the PC product user-interface. It is optimized for the 80386 and feels at least as fast on a 20mhz Compaq as APL*PLUS/UNX on a 20Mhz Sun-3. I don't know anything about their Mac product line, but one can always hope that it might someday evolve in a similar fashion. There are a lot of Mac IIs floating around here, but they are used almost exclusively for document preparation and illustration. If this is indicative of the broader market, then I imagine APL for a MAC would not be a high-volume item. Jan Prins Dept. of Computer Science UNC Chapel Hill (*) claimers and disclaimers: a trade- or service-mark is held on many of the capitalized words above by various companies; none of those companies claim that I am in any way a spokesman for them. Opinions only; no facts here.
markd@proxftl.UUCP (Mark Davidson) (12/01/88)
A question: for someone who is interested in learning about APL, is STSC's Pocket APL worth the money? Or does someone have a better suggestion? Who else besides STSC makes APL for the PC-type machines? -- In real life: Mark E. Davidson uflorida!novavax!proxftl!markd Proximity Technology Inc., 3511 NE 22nd Ave, Ft. Lauderdale FL, 33308 #define STANDARD_DISCLAIMER <Quote construction site>
wine@maui.cs.ucla.edu (David Wine) (12/02/88)
In article <1076@proxftl.UUCP> markd@proxftl.UUCP (Mark Davidson) writes: >A question: for someone who is interested in learning about APL, is STSC's >Pocket APL worth the money? Or does someone have a better suggestion? Who >else besides STSC makes APL for the PC-type machines? IBM kind of recently came out with APL2/PC. It's very nice, and I think it costs around $500. It includes user definable operators, generalized arrays, and almost everything else in APL2. You'll need a display adapter that can download fonts, or else be able to burn your own character set EPROM. --David Wine University of California at Los Angeles wine@cs.ucla.edu Computer Science Department (213) 825-6010 3531 Boelter Hall Los Angeles, CA 90024
mckee@vu-vlsi.Villanova.EDU (Bruce McKee) (12/02/88)
STSC is not the only source for Macintosh APLs. Check back issues of
MacTutor - I recall an article that listed 4-5 different products.
There was STSC, Spencer Organization, a product called MacAPL, and some
public domain product from France. The article talked about STSC's APL in the most detail, but gave references to the other entries.
Regarding: "Is APL a dying language"? I disagree.
I learned APL back in 1985, and I find it invaluable as both a prototyping
language and a way of thinking, particularly if you must map algorithms onto
parallel hardware (SIMD machines, array processors, etc.). For example, the
Connection Machine has certain operations that nicely map into APL operators.
Result: algorithms prototyped in APL can be easily re-written in parallel C
and Fortran.
APL would go much further if there were standard ASCII keywords for each
operator (I can type words faster than remembering the symbol positions).
Also needed are standard graphics and math libraries. These additions
would greatly improve APL's acceptance.
-Bruce McKee
Villanova University
>These are my opinions - your mileage may vary.ljdickey@water.waterloo.edu (Lee Dickey) (12/03/88)
In article <1076@proxftl.UUCP> markd@proxftl.UUCP (Mark Davidson) writes: >A question: for someone who is interested in learning about APL, is STSC's >Pocket APL worth the money? Or does someone have a better suggestion? Who >else besides STSC makes APL for the PC-type machines? Yes, Pocket APL is a bargain. When it was produced, it was identical to their full-blown product, with a few exceptions, like full screen editor, and some limit on the number of files one could access at one time. Two other bargains: I-APL See APL Quote Quad, vol 17/3, page 2. Sharp APL for IBM/PC See APL Quote Quad, vol 18/4, page 31. -- L. J. Dickey, Faculty of Mathematics, University of Waterloo. ljdickey@WATDCS.UWaterloo.ca ljdickey@water.BITNET ljdickey@water.UUCP ..!uunet!watmath!water!ljdickey ljdickey@water.waterloo.edu
Schizoid@cup.portal.com (FRED APPLE BONHOTAL) (12/03/88)
>APL would go farther with ASCII characterset....
I've used at least two implementations that worked with a 64-character
set (uppercase ASCII, basically); in fact, that's where I learned APL.
It wasn't until much later that I ever used the "native" APL characterset.
In one implementation, each character that wasn't represented on a "normal"
keyboard was mapped to two characters, the first of which was always $ or
@ (at the time, the "native" set didn't include $); e.g. rho was $R.
Not all of the definitions were intuitively obvious -- theta was $V,
phi was @V, del was $F, locked-del was $K -- but it wasn't that hard to
learn. It wasn't even that difficult to read. (University of Maryland)
Easier to learn, but harder to read, was the Harris implementation.
With the exception of diamond (@DI), each character was represented
by an @ followed by the first letter of the character's name followed by
the first constant after that (del was @DL, rho @RH); where there were
multiple common names, the dyadic one was given preference. (There was
one other exception, now that I think about it: delta was @DE.)
It's surprisingly useful; I wish more APL's offered this escape set.
schizoid@cup.portal.comrjfrey@kepler1.UUCP (Robert J Frey) (12/03/88)
In article <1076@proxftl.UUCP> markd@proxftl.UUCP (Mark Davidson) writes: > >...is Pocket APL worth the money?... > As a tool for learning APL, yes, IF you can upgrade it to the full APL*PLUS product later on. Also, aside from the fairly expensive 386 product, STSC's PC-based APL's don't support good things like nested arrays... ============================================================================== |Dr. Robert J. Frey | {icus, spl1, dasys1}!acsm!kepler1!rjfrey | |Kepler Financial Management, Ltd.|------------------------------------------| |100 North Country Rd., Bldg. B | The views expressed are wholly my own and| |Setauket, NY 11766 | and do not reflect those of the Indepen- | |(516) 689-6300 x.16 | dent Republic of Latvia. | ==============================================================================
Pesch@cup.portal.com (Roland Henry Pesch) (12/13/88)
Re: > APL would go much further if there were standard ASCII keywords for each > operator (I can type words faster than remembering the symbol positions). > -Bruce McKee > Villanova University Recently (APL Quote Quad v18 no 2, as I recall---I don't have it here), K.E. Iverson ---yes, the originator of the language--- published a set of keywords he recommends, as an appendix to his Dictionary of APL. I use them frequently in an (APL-written) filter to allow me to address mainframe APL from non-APL terminals, and can attest to their effectiveness. Unlike any keyword set I can remember seeing before, these were chosen with regard to their appearance, and like the primitive glyphs themselves they are well-chosen, predictable enough to recognize even when not seen before, and have mnemonic value. I have been intending to recommend their use to this newsgroup for exchanging APL algorithms; I'm glad Mr. McKee's remarks have given me the needed push. If there's sufficient interest, I will copy the table from Quote Quad and post it here (unless someone beats me to it). A short sample: the sum of the first ten integers can be named "x" with the expression x@<- +/@i 10 /Roland Pesch pesch@pa.reuter.com *or* pesch@cup.portal.com
rjfrey@kepler1.UUCP (Robert J Frey) (12/15/88)
In article <12565@cup.portal.com> Pesch@cup.portal.com (Roland Henry Pesch) writes: > >Recently...K.E. Iverson...published a set of keywords he recommends...If >there's sufficient interest, I will...post it here... > From me, there's ALOT of interest. Please post at your convenience. ============================================================================== |Dr. Robert J. Frey | {acsm, sbcs, polyof}!kepler1!rjfrey | |Kepler Financial Management, Ltd.|------------------------------------------| |100 North Country Rd., Bldg. B | The views expressed are wholly my own and| |Setauket, NY 11766 | and do not reflect those of the Indepen- | |(516) 689-6300 x.16 | dent Republic of Latvia. | ==============================================================================
metzger@mozart.uucp (Bob Metzger) (12/15/88)
Please post the recently mentioned keyword list from Ken Iverson. It is rather difficult to have discussions of APL in this group without some such convention. /Bob