IBM is abandoning APL2

[mkr]

Just noticed this today :-(

https://www-01.ibm.com/common/ssi/ShowDoc.wss?docURL=/common/ssi/rep_ca/4/899/ENUSLP21-0094/index.html&request_locale=en

It's the end of an era.

[Kerry Liles]

I had heard about that a little while back (IIRC it was on a British APL
zoom webinar) ...

Worth noting - there is a suggestion that https://log-on.com is
stepping up to provide a 'solution' for current APL2 customers...
No idea how that will go.

[Kerry Liles]

funny thing... just received the following email invitation related to
Log-On:



Invitation to a BAA Webinar
25 Mar '21 16:00 GMT

Log-On Software APL an update:
A brief explanation of what else Log-On does
and their immediate plans for Log-On APL2

Conor Hoekstra:
Conor looking for feedback ahead of his presentation
"Algorithms as a Tool of Thought" for APL Seeds on 31 March '21.

Floor will then be open for any other business.

Join the Webinar https://zoom.us/j/858532665
Passcode: 391680

See the Webinar Schedule 2021 for more information.

Join early to meet with whoever else is around. NB this meeting is recorded.

To present at future BAA Webinars please contact
eve...@britishaplassociation.org .



Kind regards,
The British APL Association

[J. Clarke]

FWIW, it looks like Log-On is now the official provider. I hope this
works out better than eComStation.

[Quadibloc]

On Wednesday, March 24, 2021 at 10:49:31 AM UTC-6, mkr wrote:
> Just noticed this today :-(
>
> https://www-01.ibm.com/common/ssi/ShowDoc.wss?docURL=/common/ssi/rep_ca/4/899/ENUSLP21-0094/index.html&request_locale=en

I made a mistake in using your URL by not expanding your post first... but I see my URL is different in any case:

https://www-01.ibm.com/common/ssi/ShowDoc.wss?docURL=/common/ssi/rep_ca/1/897/ENUS921-031/index.html&request_locale=en

John Savard

[Quadibloc]

On Wednesday, March 24, 2021 at 3:21:52 PM UTC-6, J. Clarke wrote:

> FWIW, it looks like Log-On is now the official provider. I hope this
> works out better than eComStation.

And here's the Log-On announcement:

https://log-on.com/2021/01/26/log-on-software-announces-log-on-apl2/

As the IBM announcement is no doubt the result of decreased demand for
APL, whether or not this works out will depend on whether there is still
enough demand for APL to make this business viable, as opposed to the
demand for OS/2.

John Savard

[Curtis Jones]

There should be some answers in the upcoming APL BUG meeting.

The APL Bay Area Users' Group (APL BUG), Northern California APL ACM Chapter, will meet on the 10th of May to hear Mark Schora tell about Log-On Software and its acquisition of IBM’s APL2.

Time: May 10, 2021 06:00 PM Pacific Time (US and Canada)
Join Zoom Meeting
https://acm-org.zoom.us
Meeting ID: 928 8716 0325
Passcode: ×/109 8009

Mark Schora will introduce Log-On Software and give an overview of Log-On’s recent acquisition of APL2 from IBM.

Mark is president of Log-On Software, Inc., a branch of Log-On Systems and Communication. He has over 30 years’ experience bringing solutions for the enterprise to market. Mark’s team currently focuses on delivering Log-On’s solution portfolio including APL2 to the North American marketplace. Visit
https://log-on.com/ .

[Quadibloc]

I remember that there was a discussion elsewhere where the
question "What killed APL?" came up.
The unusual character set is an obvious suspect. If a computer
doesn't have support for the APL character set, that limits the
attractiveness of APL software for that computer.
But if there were strong demand for APL, it would have been easy
enough to provide support for the APL character set, for one thing,
and for another, there's always J.
So I felt that while the unusual character set certainly *helped*,
it wasn't quite enough to do the job by itself. There would also have
to be an alternative around to lessen the demand for APL from those
who might have benefited from it.
BASIC, of course, already reduced the potential demand for APL by
a big amount, as it provided a way to quickly and easily program
computers for answers to at least some problems.
But BASIC was sort of a given - so I felt that something else in addition
to the character set _meant_ something else in addition to the character
set, given BASIC was already around, so something else besides the
two of them was needed to apply the _coup de grace_ to APL.
And I think there were two suspects.
One is the _spreadsheet_, which provides an easy way to handle
several numbers at once without even doing much programming.
Another is _Mathematica_. Here's an interactive programming
environment that does even *more* of what the science and
technology crowd is likely to want than APL could.

John Savard

[Quadibloc]

On Wednesday, April 14, 2021 at 9:28:32 PM UTC-6, Quadibloc wrote:
> and for another, there's always J.

However, while J doesn't require an APL font, and uses the ordinary
ASCII character set, it doesn't *entirely* escape from the curse of
the APL special character set. Not 100%.

And indeed, the % sign is the problem.

Because APL uses / and \ as special operators for doing things to
vectors, J, following the PDP-11 transliteration scheme for APL,
after which J is patterned to an extent, uses / and \ as they are used
in APL, and therefore uses % as the division operator.
Which, of course, users of BASIC and FORTRAN may find off-putting.
So the APL character set does cast a long enough shadow to even
have some effect on J.

John Savard

[J. Clarke]

Uh, John, what leads you to believe that APL is dead?

There are three open source versions and three commercial vendors of
whom I am aware and it is still widely used in finance.

Right now the major threat to APL is R and Python which can do
everything APL can do nearly as concisely as APL and can do a lot more
besides.

We had a project 5 years ago to transition from APL to Python. It
didn't happen and nobody is looking to commit the kind of resources
that are needed to make it happen.

[Quadibloc]

On Wednesday, April 14, 2021 at 11:45:07 PM UTC-6, J. Clarke wrote:

> Uh, John, what leads you to believe that APL is dead?

It does depend on what you mean by "dead". It's just under the radar...

> There are three open source versions and three commercial vendors of
> whom I am aware and it is still widely used in finance.

> Right now the major threat to APL is R and Python which can do
> everything APL can do nearly as concisely as APL and can do a lot more
> besides.

You have a point... but at the moment, APL is sufficiently far under the radar
that while Python is certainly a popular and useful language, I wouldn't think
of it... _in connection_ with APL enough to think of it as being a threat to
APL. After all, there will always be alternative programming languages.

And the _reason_ why languages like R and Python are being invented to
provide some of the same facilities as APL is now no longer because
people are being inspired by APL to do what APL did and more besides...
but because, not having _heard_ of APL, they're independently seeing a need
to do some of the same kinds of things.

I may be exaggerating; it's certainly possible that Guido van Rossum is
knowledgeable enough about the history of computing that he not only
knows what APL is, but he could even talk about MATH-MATIC and
FLOW-MATIC and the Klerer-May system.

Mathematica and the spreadsheet helped to define the climate in the
relatively early days of the personal computer - VisiCalc dating from
the Apple II, and Mathematica from the Macintosh.

IBM included BASIC with the IBM PC; APL was also provided later,
as software that cost extra and which required the 8087 co-processor.
If APL had been "a thing" to the extent of thriving as it did in the 1970s,
it would have had to have been to be expected that keyboards would have
the APL character set on them, and text displays would include the APL
character set as an alternate mode of operation.

There was the IBM 5100. There was the MCM/70. There was the VideoBrain
Family Computer with APL/S, based on the Fairchild F8 microprocessor.

Of course, though, pointing to the spreadsheet, or even APL's character set,
is really looking at secondary causes.

The _primary_ cause is clear; when personal computers *first* came out,
BASIC was the programming language that could have sufficiently small
implementations to be included with them.

Later on, in the age of the Macintosh and Windows 3.1, computers didn't
come with programming languages period, so by the time they were ready
to have Fortran compilers or APL interpreters, personal computers, as
a consumer product, didn't include programming languages.

You could, of course, _get_ programming languages for them, and C was
what was typically used to develop for microcomputers.

You go to the store and buy a computer... it will have a graphical user
interface, which a mini or mainframe from the 1970s would not have
had, but it won't have a Fortran compiler, a COBOL compiler, an assembler,
an APL interpreter... or even a BASIC interpreter.

With Linux, one can indeed turn a personal computer into an approximation
of a mainframe computer of yore (and then there's the mighty Hercules)...
but that's just not what they're _for_.

John Savard

[Quadibloc]

On Thursday, April 15, 2021 at 12:34:23 AM UTC-6, Quadibloc wrote:

> There was the IBM 5100. There was the MCM/70. There was the VideoBrain
> Family Computer with APL/S, based on the Fairchild F8 microprocessor.

I forgot what was perhaps the best possible example of what the
microcomputer "might have been"... the Commodore SuperPET.

It included a 6809 processor in addition to the standard 6502, and that gave
it enough power to be provided not only with Waterloo MicroAPL, but their
other languages, such as Fortran. So even before the transition to 16 bits,
there was one computer, unfortunately an expensive one, that took the
microcomputer in a "big computer" direction.

Of course, there were APL interpreters for other computers as well, including
the Amiga, the Atari ST, and the Sinclair QL. You could get FORTRAN for the
Commodore 128 for that matter...

John Savard

[J. Clarke]

On Wed, 14 Apr 2021 23:34:22 -0700 (PDT), Quadibloc
<jsa...@ecn.ab.ca> wrote:

>On Wednesday, April 14, 2021 at 11:45:07 PM UTC-6, J. Clarke wrote:
>
>> Uh, John, what leads you to believe that APL is dead?
>
>It does depend on what you mean by "dead". It's just under the radar...
>
>> There are three open source versions and three commercial vendors of
>> whom I am aware and it is still widely used in finance.
>
>> Right now the major threat to APL is R and Python which can do
>> everything APL can do nearly as concisely as APL and can do a lot more
>> besides.
>
>You have a point... but at the moment, APL is sufficiently far under the radar
>that while Python is certainly a popular and useful language, I wouldn't think
>of it... _in connection_ with APL enough to think of it as being a threat to
>APL. After all, there will always be alternative programming languages.

Well that's likely because you've never worked in insurance. The old
actuaries at our company who know a programming language at all know
APL. The young ones know Python. They do the same sorts of things
with them. The Society of Actuaries complicates things a bit by
requiring all new actuaries to know R.

In that line of work, APL, R, and Python are the Big 3.

>And the _reason_ why languages like R and Python are being invented to
>provide some of the same facilities as APL is now no longer because
>people are being inspired by APL to do what APL did and more besides...
>but because, not having _heard_ of APL, they're independently seeing a need
>to do some of the same kinds of things.

R is not particularly new--it dates to 1976. It didn't become popular
until an open-source version became available in the late '90s, right
about the time that Python was hitting. The same didn't happen for
APL--I'm not sure why--possibly because gnu APL and A+ are not
particularly friendly and NARS2000 is Windows-only.

>I may be exaggerating; it's certainly possible that Guido van Rossum is
>knowledgeable enough about the history of computing that he not only
>knows what APL is, but he could even talk about MATH-MATIC and
>FLOW-MATIC and the Klerer-May system.
>
>Mathematica and the spreadsheet helped to define the climate in the
>relatively early days of the personal computer - VisiCalc dating from
>the Apple II, and Mathematica from the Macintosh.

The spreadsheet certainly but Mathematica is in a different direction.

>IBM included BASIC with the IBM PC; APL was also provided later,
>as software that cost extra and which required the 8087 co-processor.
>If APL had been "a thing" to the extent of thriving as it did in the 1970s,
>it would have had to have been to be expected that keyboards would have
>the APL character set on them, and text displays would include the APL
>character set as an alternate mode of operation.

I really wish that people would get over their fixation with
keyboards. NOBODY I know uses a special keyboard. Mine right now is
a Logitech gamer board. The one on my laptop works fine with it. So
does the "official" Raspberry Pi keyboard (and note that APL is free
for personal use on the Pi--so is Mathematica).

And it's 2021--the only text display that is relevant to APL at this
point is the 3270, and it has the APL character set.

>There was the IBM 5100. There was the MCM/70. There was the VideoBrain
>Family Computer with APL/S, based on the Fairchild F8 microprocessor.

The 5100 and MCM were both rather costly. The Videobrain seems to
have had a variety of problems.

>Of course, though, pointing to the spreadsheet, or even APL's character set,
>is really looking at secondary causes.
>
>The _primary_ cause is clear; when personal computers *first* came out,
>BASIC was the programming language that could have sufficiently small
>implementations to be included with them.

By that logic BASIC would have killed Fortran, COBOL, and C. It
didn't.

>Later on, in the age of the Macintosh and Windows 3.1, computers didn't
>come with programming languages period, so by the time they were ready
>to have Fortran compilers or APL interpreters, personal computers, as
>a consumer product, didn't include programming languages.
>
>You could, of course, _get_ programming languages for them, and C was
>what was typically used to develop for microcomputers.
>
>You go to the store and buy a computer... it will have a graphical user
>interface, which a mini or mainframe from the 1970s would not have
>had, but it won't have a Fortran compiler, a COBOL compiler, an assembler,
>an APL interpreter... or even a BASIC interpreter.
>
>With Linux, one can indeed turn a personal computer into an approximation
>of a mainframe computer of yore (and then there's the mighty Hercules)...
>but that's just not what they're _for_.

What's your point? I have C, C++, BASIC in two flavors, APL, Python,
R, and Powershell on my work machine, plus C, FORTRAN, COBOL and REXX
on the mainframe. The days where a programmer works in one language
for his entire career or even his entire work day are pretty much
over. You use the right tool for the job. Most of the time that's
APL, BASIC, or Python, and to perform a task I may have pieces in all
three of them.

[A Ask]

J Clarke: "Uh, John, what leads you to believe that APL is dead?"
Me: What leads you to believe that APL is NOT dead?

If I needed to use a language to get through a phase, say Actuarial examinations, I would happily use APL.
If I wanted a career in Software Development, I would NOT start by learning APL (no up-to-date references, dearth of worked examples and what exists is hard to find,).

I love APL as the tool that gets the job done WITHOUT the intrusion of usual programming considerations. However, I find MicroAPL & IBM leaving the APL market rather disturbing; it does not bode well for the future of APL.

[DaveW]

I wish to point out that the first spreadsheet was APLDOT, written in APL. It is not universally recognized as the first, because even though Bob Jernigan created it in the early 70s (to analyze the break-up of the Penn Central RR for DoT), he did not publish about it for another decade.
Dave

[J. Clarke]

On Thu, 15 Apr 2021 07:10:12 -0700 (PDT), A Ask
<AJAY_A...@yahoo.co.uk> wrote:

>J Clarke: "Uh, John, what leads you to believe that APL is dead?"
>Me: What leads you to believe that APL is NOT dead?

Well, for one thing where I work there are more APL developers than I
can find, and management has been trying to kill it for years without
success. I understand that this is not atypical in finance.

>If I needed to use a language to get through a phase, say Actuarial examinations, I would happily use APL.
>If I wanted a career in Software Development, I would NOT start by learning APL (no up-to-date references, dearth of worked examples and what exists is hard to find,).

I wouldn't pick R or Visual Basic for Applications for a career in
software development either. There is more world out there than
career software developers.

>I love APL as the tool that gets the job done WITHOUT the intrusion of usual programming considerations. However, I find MicroAPL & IBM leaving the APL market rather disturbing; it does not bode well for the future of APL.

On the other hand Dyalog seems to be doing well and is aggressively
marketing. I tried to buy a copy of APL2 a while back and after
playing phone tag with IBM and then IBM's "approved vendor" I finally
gave up trying.

The hard part is getting people to try it IMO. Once they've used it
for a bit they get hooked.

[Quadibloc]

Computer languages... exist.

My point was that the world of computing has moved on from the days
before the microprocessor. So programming languages don't have the
central role for computers that they did in, say, 1972.

After all, there was an APL for the CDC 6600. There was an APL for the
SDS 940. There was an APL for the PDP-11. At one time, it was felt that
if you were making a computer, you had to offer APL for it. Not any more!

Certainly, if you have an x86-compatible running Windows 10, you can
*get* APL for it. I'm aware of APL X and the turnkey APL/360 version of
MVT on Hercules, for example.

However, there is a difference - a _big_ difference - between that and a
situation like this:

Today, Commodore announced their new Commodore 128 computer which
would be available in three models: one with an additional Z-80 processor, for which
CP/M would be available, one with an additional 6809 processor, for which the same
Waterloo software suite as offered with the SuperPET would be available, and one
which had both the Z-80 and the 6809, and the availablity of both sets of software.

Or one where the Commodore Amiga and/or the Atari ST came with APL in
addition to BASIC.

The difference?

The existence of a large, vibrant user community for the language.

The fact that IBM, Kenneth E. Iverson's former employer, and the company
that first brought APL to the world in the form of APL\360, is throwing in the
towel... is not nothing.

Maybe it doesn't qualify as earth-shaking simply because APL didn't fall from
having a large user community to this in one day. Is it really surprising, even,
that IBM has finally decided to stop keeping a niche product on life-support?
But it's definitely a milestone.

John Savard

[Quadibloc]

On Thursday, April 15, 2021 at 10:56:17 AM UTC-6, J. Clarke wrote:

> Well, for one thing where I work there are more APL developers than I
> can find, and management has been trying to kill it for years without
> success. I understand that this is not atypical in finance.

Why would management be trying to kill it, if it's working well?

I'm not saying it isn't working well. Instead, we all know what the
reason is. They're worried that, because there aren't enough shops
using APL - in other sectors besides finance, preferably - one of
these days, the bottom is going to fall out from under vendor
support.

> The hard part is getting people to try it IMO. Once they've used it
> for a bit they get hooked.

Hey, we're all agreed that APL is a wonderful language.

Now that APL2 is no longer an IBM product, though, APL has lost
its last remaining shred of visibility. Of course it's harder to get
people to try something they've never heard of!

And, strange to relate, there are actually people old enough to
shave these days who don't remember when you usually used
a computer by means of punched cards, but had to go to a 2741
time-sharing terminal for APL.

In fact, there are even adults these days who never lived through the
days when 8-bit microcomputers were a thing. Someone born in
1981, the year IBM announced the IBM Personal Computer, which
transitioned the industry to 16 bits, would be around 40 years old now.

I know it's hard to even imagine the perspective of a 30-year-old,
for whom computers for direct use (as opposed to embedded
processors) always had 32-bit processors and graphical user
interfaces...

There is Tcl/Tk. There was Clarion. While a revival of APL would be
nice, in my opinion, the most single glaring deficiency of today's world
of computers is this:

Back in the days of command-line interfaces, with BASIC,
nearly anyone could learn to write a computer program.

It might take some time before one could write one with a professional
polish, true.

Today, though, the learning curve for writing a Windows application
is a very steep one. You can't really just write a program at
a keyboard - you will also need graphical screen builders.
And the part you do write at the keyboard will involve
insanely complex things like the Microsoft Foundation
Classes.

So no wonder writing programs is left to the large companies that
can hire large teams of programmers to develop them.

John Savard

[Quadibloc]

On Thursday, April 15, 2021 at 5:17:27 AM UTC-6, J. Clarke wrote:

> I really wish that people would get over their fixation with
> keyboards. NOBODY I know uses a special keyboard. Mine right now is
> a Logitech gamer board. The one on my laptop works fine with it. So
> does the "official" Raspberry Pi keyboard (and note that APL is free
> for personal use on the Pi--so is Mathematica).

> And it's 2021--the only text display that is relevant to APL at this
> point is the 3270, and it has the APL character set.

Oh, you mean like this...

http://www.quadibloc.com/comp/images/3270apl2.gif

(from the page
http://www.quadibloc.com/comp/kyb01.htm
)

John Savard

[J. Clarke]

No, I don't mean like that.

Again with the keyboard. Grok the concept: NO APL USER CARES ABOUT
THE KEYBOARD.

It is the display character set that matters, not the pictures on the
keyboard that nobody ever looks at except hunt-and-peck typists.

[J. Clarke]

On Thu, 15 Apr 2021 10:31:58 -0700 (PDT), Quadibloc
<jsa...@ecn.ab.ca> wrote:

>On Thursday, April 15, 2021 at 10:56:17 AM UTC-6, J. Clarke wrote:
>
>> Well, for one thing where I work there are more APL developers than I
>> can find, and management has been trying to kill it for years without
>> success. I understand that this is not atypical in finance.
>
>Why would management be trying to kill it, if it's working well?

In part because a while back some loon wrote a book in which they
railed against "the cult of APL". And in part because they want to be
"cool kids" and "the cool kids" don't use "klunky old languages".
Hence we have a process whereby we query an admin system to obtain a
list of information, convert the binary to a CSV, download the CSV to
a LAN file share, upload the CSV to the AWS cloud, import the CSV into
a table in an SQL database, extract the table, download it to a
network share, upload it to the mainframe, convert it to a binary, run
a process on it, convert the output from that process to a CSV,
download, copy to AWS, load into SQL table, download from SQL table,
copy back up to mainframe, convert to binary, and then perform the
next step in the process, all so that we can have stuff "in the cloud"
like the cool kids;.

>I'm not saying it isn't working well. Instead, we all know what the
>reason is. They're worried that, because there aren't enough shops
>using APL - in other sectors besides finance, preferably - one of
>these days, the bottom is going to fall out from under vendor
>support.

No, that's what I'm worried about. Management has no idea how many
shops are using APL, or any of the rest. Hell, the last time I talked
to one of them about it he thought I was talking about something to do
with Amazon.

>> The hard part is getting people to try it IMO. Once they've used it
>> for a bit they get hooked.
>
>Hey, we're all agreed that APL is a wonderful language.
>
>Now that APL2 is no longer an IBM product, though, APL has lost
>its last remaining shred of visibility. Of course it's harder to get
>people to try something they've never heard of!

Hardly. APL2 hasn't been particularly visible for a long time. Dyalog
is the big player.

>And, strange to relate, there are actually people old enough to
>shave these days who don't remember when you usually used
>a computer by means of punched cards, but had to go to a 2741
>time-sharing terminal for APL.

So? None of the APL developers I work with have ever seen a 2741.

>In fact, there are even adults these days who never lived through the
>days when 8-bit microcomputers were a thing. Someone born in
>1981, the year IBM announced the IBM Personal Computer, which
>transitioned the industry to 16 bits, would be around 40 years old now.

You still aren't making any kind of point.

>I know it's hard to even imagine the perspective of a 30-year-old,
>for whom computers for direct use (as opposed to embedded
>processors) always had 32-bit processors and graphical user
>interfaces...
>
>There is Tcl/Tk. There was Clarion. While a revival of APL would be
>nice, in my opinion, the most single glaring deficiency of today's world
>of computers is this:
>
>Back in the days of command-line interfaces, with BASIC,
>nearly anyone could learn to write a computer program.

So? You still aren't making a point.

>It might take some time before one could write one with a professional
>polish, true.
>
>Today, though, the learning curve for writing a Windows application
>is a very steep one. You can't really just write a program at
>a keyboard - you will also need graphical screen builders.
>And the part you do write at the keyboard will involve
>insanely complex things like the Microsoft Foundation
>Classes.
>
>So no wonder writing programs is left to the large companies that
>can hire large teams of programmers to develop them.

What leads you to believe that every program needs a GUI? Geez, talk
about making mountains out of molehills.

[Quadibloc]

On Thursday, April 15, 2021 at 1:50:08 PM UTC-6, J. Clarke wrote:
> And in part because they want to be
> "cool kids" and "the cool kids" don't use "klunky old languages".

To phrase this in a way which makes visible the _rational_ objection: it
will be easier to recruit staff if the position provides experience in
languages that are currently in demand.

> What leads you to believe that every program needs a GUI? Geez, talk
> about making mountains out of molehills.

It's surprising to me that the point I was making is not clear and
obvious.

Back in the command-line era...

Command-line programs could be written in a simple and natural
manner by anyone who was trained in a compiled language.

Thus, if you were a programmer, you could write applications
programs.

Today, though, with the prevalence of the GUI - *and* the way
operating systems are designed to support the GUI - there is no
longer a simple path from "learn how to program" and "write an
application that at least approaches those which are commercially
sold and distributed".

Now the tools used to build applications are much more complex
than a compiler for a programming language.

Of course, though, what with APL not normally producing
compiled executables, this is not so much of an issue for APL
specifically.

John Savard

[Quadibloc]

On Thursday, April 15, 2021 at 1:32:22 PM UTC-6, J. Clarke wrote:

> Again with the keyboard. Grok the concept: NO APL USER CARES ABOUT
> THE KEYBOARD.

> It is the display character set that matters, not the pictures on the
> keyboard that nobody ever looks at except hunt-and-peck typists.

1) To be able to touch-type in _APL_ would require that one has had quite
a bit of experience in it, and uses it a lot.

2) For characters to reach the display, the keyboard layout must
generate them, whatever may be printed on the keys.

3) But the keyboard is more a consequence than a cause. Back when
APL was actually popular, there was a version of the DECwriter,
there was a version of the Tektronix 4010, and so on and so forth,
with an APL keyboard. If you don't see APL characters on the keyboards
of computers today, that's a sign it isn't being used as much.

John Savard

[Quadibloc]

On Thursday, April 15, 2021 at 1:50:08 PM UTC-6, J. Clarke wrote:
> On Thu, 15 Apr 2021 10:31:58 -0700 (PDT), Quadibloc
> <jsa...@ecn.ab.ca> wrote:

> >Now that APL2 is no longer an IBM product, though, APL has lost
> >its last remaining shred of visibility. Of course it's harder to get
> >people to try something they've never heard of!

> Hardly. APL2 hasn't been particularly visible for a long time. Dyalog
> is the big player.

That may be.

But surely you can see the problem.

Dyalog may be the foremost company in the APL field. But it isn't Microsoft,
it isn't Apple, and it isn't IBM. Nor is it Google or even Samsung., Intel, or Nvidia.
Or AMD.

Still, you may be quite right that I am... exaggerating. There's a big difference
between "dead" and "not totally in your face". Of course APL can survive for
decades more, languishing in relative obscurity.

John Savard

[J. Clarke]

On Thu, 15 Apr 2021 20:02:18 -0700 (PDT), Quadibloc
<jsa...@ecn.ab.ca> wrote:

>On Thursday, April 15, 2021 at 1:32:22 PM UTC-6, J. Clarke wrote:
>
>> Again with the keyboard. Grok the concept: NO APL USER CARES ABOUT
>> THE KEYBOARD.
>
>> It is the display character set that matters, not the pictures on the
>> keyboard that nobody ever looks at except hunt-and-peck typists.
>
>1) To be able to touch-type in _APL_ would require that one has had quite
>a bit of experience in it, and uses it a lot.

Everybody at our company does fine with a piece of paper pinned to the
wall in their cubicle.

>2) For characters to reach the display, the keyboard layout must
>generate them, whatever may be printed on the keys.

All it has to do is generate a signal indicating that the key in a
particular location was pressed. Converting those signals to
character encodings is done in software on the computer. From the
viewpoint of a modern computer, APL is just another foreign language,
there isn't any more magic involved than in making glyphs for Cyrillic
or Arabic.

>3) But the keyboard is more a consequence than a cause. Back when
>APL was actually popular, there was a version of the DECwriter,
>there was a version of the Tektronix 4010, and so on and so forth,
>with an APL keyboard. If you don't see APL characters on the keyboards
>of computers today, that's a sign it isn't being used as much.

That's like saying "if you don't see Chinese characters on the

keyboards of computers today, that's a sign it isn't being used as
much".

I have two APL keyboards. I don't use either because the glyphs
marked on the keys do not match the mappings for the version of APL
that I use, and I find them more confusing than helpful.

[J. Clarke]

On Thu, 15 Apr 2021 19:47:39 -0700 (PDT), Quadibloc
<jsa...@ecn.ab.ca> wrote:

>On Thursday, April 15, 2021 at 1:50:08 PM UTC-6, J. Clarke wrote:
>> And in part because they want to be
>> "cool kids" and "the cool kids" don't use "klunky old languages".
>
>To phrase this in a way which makes visible the _rational_ objection: it
>will be easier to recruit staff if the position provides experience in
>languages that are currently in demand.

By that logic COBOL is dead.

Most APL users are not programmers. My boss did me the kindness of
getting my title changed from "Quant" to "Developer". Most of the
people I work with have titles of "Quant" or "Actuary". You seem to
be living in the IT ghetto.

>> What leads you to believe that every program needs a GUI? Geez, talk
>> about making mountains out of molehills.
>
>It's surprising to me that the point I was making is not clear and
>obvious.
>
>Back in the command-line era...
>
>Command-line programs could be written in a simple and natural
>manner by anyone who was trained in a compiled language.
>
>Thus, if you were a programmer, you could write applications
>programs.
>
>Today, though, with the prevalence of the GUI - *and* the way
>operating systems are designed to support the GUI - there is no
>longer a simple path from "learn how to program" and "write an
>application that at least approaches those which are commercially
>sold and distributed".
>
>Now the tools used to build applications are much more complex
>than a compiler for a programming language.
>
>Of course, though, what with APL not normally producing
>compiled executables, this is not so much of an issue for APL
>specifically.

Once again you seem to be falling into the IT ghetto. I need numbers
crunched. I don't need a pretty GUI. My colleagues need numbers
crunched. They don't need pretty GUIs. You are assuming that the
only utility of a tool is to produce commercial software. The fact is
that producing GUIs in APL isn't any more painful than producing them
in Python, so by your logic Python must be dead.

[J. Clarke]

On Fri, 16 Apr 2021 07:25:52 -0700 (PDT), Quadibloc

<jsa...@ecn.ab.ca> wrote:

>On Thursday, April 15, 2021 at 1:50:08 PM UTC-6, J. Clarke wrote:
>> On Thu, 15 Apr 2021 10:31:58 -0700 (PDT), Quadibloc
>> <jsa...@ecn.ab.ca> wrote:
>
>> >Now that APL2 is no longer an IBM product, though, APL has lost
>> >its last remaining shred of visibility. Of course it's harder to get
>> >people to try something they've never heard of!
>
>> Hardly. APL2 hasn't been particularly visible for a long time. Dyalog
>> is the big player.
>
>That may be.
>
>But surely you can see the problem.
>
>Dyalog may be the foremost company in the APL field. But it isn't Microsoft,
>it isn't Apple, and it isn't IBM. Nor is it Google or even Samsung., Intel, or Nvidia.
>Or AMD.

Geez, Ford is no Tesla. So what?

[Quadibloc]

On Friday, April 16, 2021 at 9:11:41 AM UTC-6, J. Clarke wrote:

> That's like saying "if you don't see Chinese characters on the
> keyboards of computers today, that's a sign it isn't being used as
> much".

Well, it is at least a sign that the ChangJie input method isn't
used as much as the Pinyin input method, if you're talking about
keyboards in places where Chinese is commonly spoken.

John Savard

[Quadibloc]

On Friday, April 16, 2021 at 9:19:31 AM UTC-6, J. Clarke wrote:
> On Fri, 16 Apr 2021 07:25:52 -0700 (PDT), Quadibloc
> <jsa...@ecn.ab.ca> wrote:

> >On Thursday, April 15, 2021 at 1:50:08 PM UTC-6, J. Clarke wrote:

> >Dyalog may be the foremost company in the APL field. But it isn't Microsoft,
> >it isn't Apple, and it isn't IBM. Nor is it Google or even Samsung., Intel, or Nvidia.
> >Or AMD.

> Geez, Ford is no Tesla. So what?

I would have put it that Tesla is no Ford (or GM or AMC... unless AMC isn't still
around). So perhaps part of my problem is that I'm a dinosaur...

John Savard

[Quadibloc]

On Friday, April 16, 2021 at 9:16:50 AM UTC-6, J. Clarke wrote:
> On Thu, 15 Apr 2021 19:47:39 -0700 (PDT), Quadibloc
> <jsa...@ecn.ab.ca> wrote:

> >Of course, though, what with APL not normally producing
> >compiled executables, this is not so much of an issue for APL
> >specifically.

> Once again you seem to be falling into the IT ghetto. I need numbers
> crunched. I don't need a pretty GUI.

And people who need numbers crunched have plenty of tools... spreadsheets,
Mathematica, and, yes, even Python. APL now has more competition than it
used to have in this area, which has forced it into a narrower niche.

John Savard

[Quadibloc]

On Friday, April 16, 2021 at 9:11:41 AM UTC-6, J. Clarke wrote:

> I have two APL keyboards. I don't use either because the glyphs
> marked on the keys do not match the mappings for the version of APL
> that I use, and I find them more confusing than helpful.

And you don't think that's a sign that APL is declining in popularity?

John Savard

[J. Clarke]

Nope, I think it's a sign that every APL vendor has a different
keyboard layout and you can't make one keyboard that fits all of them.
Hell, the vendor whose product I use for work has _two_ APL keyboard
layouts and I use a different one from the rest of the team.

[J. Clarke]

On Sat, 17 Apr 2021 06:16:14 -0700 (PDT), Quadibloc

<jsa...@ecn.ab.ca> wrote:

>On Friday, April 16, 2021 at 9:16:50 AM UTC-6, J. Clarke wrote:
>> On Thu, 15 Apr 2021 19:47:39 -0700 (PDT), Quadibloc
>> <jsa...@ecn.ab.ca> wrote:
>
>> >Of course, though, what with APL not normally producing
>> >compiled executables, this is not so much of an issue for APL
>> >specifically.
>
>> Once again you seem to be falling into the IT ghetto. I need numbers
>> crunched. I don't need a pretty GUI.
>
>And people who need numbers crunched have plenty of tools... spreadsheets,
>Mathematica, and, yes, even Python. APL now has more competition than it
>used to have in this area, which has forced it into a narrower niche.

You've never actually done much with spreadsheets if you think that
spreadsheets are any kind of substitute for APL. Load a half-gigabyte
table into Excel and see what it does.

Nobody I know uses Mathematica. If you think it's competition for any
programming language you have never used it.

Python and R are the real competition but they will only "kill APL"
when there is a tool that will reliably and automatically convert an
APL workspace into a Python module.

APL is like COBOL--there's too much code out there.

[J. Clarke]

On Sat, 17 Apr 2021 06:13:35 -0700 (PDT), Quadibloc

<jsa...@ecn.ab.ca> wrote:

>On Friday, April 16, 2021 at 9:19:31 AM UTC-6, J. Clarke wrote:
>> On Fri, 16 Apr 2021 07:25:52 -0700 (PDT), Quadibloc
>> <jsa...@ecn.ab.ca> wrote:
>
>> >On Thursday, April 15, 2021 at 1:50:08 PM UTC-6, J. Clarke wrote:
>
>> >Dyalog may be the foremost company in the APL field. But it isn't Microsoft,
>> >it isn't Apple, and it isn't IBM. Nor is it Google or even Samsung., Intel, or Nvidia.
>> >Or AMD.
>
>> Geez, Ford is no Tesla. So what?
>
>I would have put it that Tesla is no Ford (or GM or AMC... unless AMC isn't still
>around). So perhaps part of my problem is that I'm a dinosaur...

Just for the record, Ford market cap 47.95 billion, Tesla market cap
710.08 billion. Tesla at this point looks to the automotive industry
like Godzilla rising from the sea.

AMC's been dead 30 years.