Definitions are important or else words are subject to widely varying
interpretations. 'dying' appears to be equated to 'extinct' in these
claims. I suggest that as long as someone needs to understand the
underlying architecture of a computer, assembler language will be a
useful skill for them - as a learning tool, if nothing else. It may not
be a good choice for a college to create a curriculum, but certainly a
few might be able to offer a class or two and have 20-50 students in it
each year.
In 1991 Stewart Alsop infamously predicted the end of mainframes by
1996. He seemed to be very certain, based on the trends he was
observing. In other words, at that time mainframes were 'dying'. In
the mid-2000s the developers I worked with (whose primary coding
language was assembler) were very paranoid that their jobs would be
outsourced to India, because so many other mainframe jobs had been
(operations, application development, etc.). My observation was that
this would be very unlikely because India had no incentive to develop an
education infrastructure that might be useful for only a few thousand
students (at most). Instead, by focusing on COBOL, C, Java, HTML, CSS,
Javascript, etc. they could educate and provide useful (saleable) skills
to many millions.
There have been more than a few threads on this list about how
mainframes are dying as fewer 'new' installations are identified.
Those debates have been just as vehement. The reality seems to be that
mainframes are here to stay for a long time, and therefore HLASM skills
will be necessary for a long time for at least a few folks. I'm sure
IBM will make sure that internally there are assembler experts educated
and employed inside the company. The same is likely true for vendors of
other hardware/operating systems.
As for AI taking over that role, I am really skeptical. I am doing some
education on machine learning and what it involves. The term implies the
AI gets feedback regarding the accuracy of its
conclusions/predictions/etc. - presumably from humans. If a human is
not able to verify the accuracy of the results, how does the machine
learn? You can ask AI to 'write an assembler program' to do something,
but how do you know it works? Either you test it and simply say "no,
that's not it" (which is almost useless for machine learning) or you
have a subject matter expert analyze it and provide a correction. At
what point do you no longer need a subject matter expert? When every
problem has been solved? If every problem has not yet been solved,
doesn't someone need to verify the results of any AI generated solution,
and provide useful feedback on what needs to be changed?
Currently, one can find a great many examples of coding solutions in
popular languages in various easily searched open source repositories,
and with many millions with expertise able to provide feedback AI can
learn very quickly. Most assembler solutions are proprietary, and
therefore not searchable. Even if you were to search CBT online you
wouldn't find a lot of details until you download, extract and then
analyze the code. Even then, detailed descriptions of what a
particular section of code is doing and how each piece relates to a
specific action (DCB vs OPEN/CLOSE vs GET/PUT, etc.) would take a bit
more analysis and feedback. Could AI do it? Sure, eventually - when
experts provide the feedback necessary for an AI to 'learn' what is and
is not correct, but with relatively few of them in the world and most of
those having little to no incentive to 'teach' the AI proprietary
information that the 'unwashed masses' could utilize freely and easily,
I don't see it happening for many years.
My conclusion is that those with current assembler skills are safe for
as long as they want to practice it professionally - including those
just starting their careers. I certainly wouldn't recommend to
everyone as a career objective - but a few will find a career, and the
rest may find that the understanding they gain of the underlying
architecture and hardware will be useful in finding other professional
opportunities... including writing better code in a higher level language.
My final suggestion to people is: If you don't like (consistently) what
someone writes, don't reply. If you can't resist, put them on
ignore. Once upon a time I felt it necessary to respond to every
single reply and retort, then I matured.
I now return you to your regular programming...
On 9/17/2023 7:53 AM, Bill Johnson wrote:
Which of my facts aren’t true? Assembler is dying. Confirmed by one of
Assembler’s experts, Ray Mullins.
With a name like Bill Johnson, I would either be Scandinavian or black most
likely. I’m the former. Swedish mostly. Are you aware of how Swedes are named?
Here’s some context easily found on the web on highly credible sites.
First names often provide information such as gender and historical trends,
cultural backgrounds, and nationality. Last names provide information on the
roots of the family system and the origin of ethnicity. Using both can increase
the chances of being able to predict someone's ethnicity with greater
confidence.
Sent from Yahoo Mail for iPhone
On Sunday, September 17, 2023, 10:30 AM, Seymour J Metz <sme...@gmu.edu> wrote:
I have a talent for facts.
ObSeanSpicer Alternative facts.
anyone can tell with near certainty the color and ethnicity of
most posters here simply based on their names.
It's unfortunate that being certain and being correct are two very different
things. FWIW, I would have guessed from your name that you were of Anglo-Saxon
ancestry, but you claim otherwise.
________________________________________
From: IBM Mainframe Discussion List <IBM-MAIN@LISTSERV.UA.EDU> on behalf of Bill
Johnson <00000047540adefe-dmarc-requ...@listserv.ua.edu>
Sent: Saturday, September 16, 2023 10:38 AM
To: IBM-MAIN@LISTSERV.UA.EDU
Subject: Re: AI will surpass human intelligence!
I have a talent for facts. Which I’ve posted throughout this diatribe. The
problem is many of you don’t like someone outside the cult showing you up so
you attack and make ridiculous accusations and idiotic comparisons like the
college one.
Fact - assembler is dying a slow tortuous death.
Fact - anyone can tell with near certainty the color and ethnicity of most
posters here simply based on their names.
Sent from Yahoo Mail for iPhone
On Saturday, September 16, 2023, 1:09 AM, g...@gabegold.com <g...@gabegold.com>
wrote:
You could also incorrectly "tell" that I didn't have a college degree.
You have a talent for gratuitous insults and an obsession with starting and
then arguing about wild digressions from whatever was the topic at hand.
AI to assembler language to system programming demographics is quite the
meander; each of those would have been worth separate discussions, but not in
the unpleasant way you bring them up and then harangue anyone who disagrees
with you.
A basic question is why, since you disdain everyone here and disagree with most
of what people say, you're still here. You're not changing minds, not making
friends, not providing useful information, and not contributing anything.
You're excellent at trolling though, which seems to be your sad reason for
being. So you're entertaining, in a ghastly sort of way.
On Fri, 15 Sep 2023 18:33:25 +0000, Bill Johnson <mellonb...@yahoo.com> wrote:
Are you that naive? Ever been to Share? The current makeup of mainframe systems
programmers is 65% white, 15% Asian, 9% Hispanic & 5% African American.
Factoring in most of the heavy posters are over 50, the demographics are even more
white since almost all systems programmers who got into IT in the 50’s through the
70’s were almost exclusively white. Guess what? I can also tell which people are
Jewish, Asian, Hispanic, Scandinavian, Irish, and most other ethnicities. That’s
not nearly as easy but it’s not as difficult as you’d think. I’d bet Oujeski isn’t
a black guy. My Father in Law’s name was Majewski.
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
