State of New Jersey needs COBOL programmers

Jay Jaeger cube1 at charter.net
Sun Apr 5 21:54:31 CDT 2020


On 4/5/2020 8:46 PM, ben via cctalk wrote:
> On 4/5/2020 6:44 PM, Toby Thain via cctalk wrote:
> 
>>> Q: If Bill Gates hadn't written a BASIC interpreter, where would we
>>> be now?
>>>
>>>
>>
>> Sounds like you've never heard of Lisp.
>>
> 
> All the old programmers speak with a LISP.
> I view computer science ... teaching is what 'trending now' since
> schools promote grinding out students, not teaching real world problems.
> As for Gates, the right place at the right time, and having access to a
> main frame.

On 4/5/2020 8:46 PM, ben via cctalk wrote:> On 4/5/2020 6:44 PM, Toby
Thain via cctalk wrote:
>
>>> Q: If Bill Gates hadn't written a BASIC interpreter, where would we
>>> be now?
>>>
>>>
>>
>> Sounds like you've never heard of Lisp.
>>
>
> All the old programmers speak with a LISP.
> I view computer science ... teaching is what 'trending now' since
> schools promote grinding out students, not teaching real world problems.
> As for Gates, the right place at the right time, and having access to a
> main frame.
Well, I'd have called the PDP-10 a minicomputer nonetheless, but that's
just me.  ;)  Plus, he was essentially stealing campus resources for
personal aggrandizement.  Its one thing to mess around learning stuff
outside of a course setting.  Quite another to take that time, without
even asking, in order to turn a profit.  Not a fan.

As for Lots of Insipid Stupid Parentheses - I can't see that as a
beginning language.  For anyone.  Of the many many (well over a hundred)
programmers I worked with outside of a university setting, less than a
handful had even *heard* of it.

<RANT>

As for teaching, there are multiple needs.  One size doesn't fit all
institutions.   One is plain old programming.  That does get taught in
computer science at universities, but it really ought to be just a means
to an end: learning about how programs get written, how to compile them,
data structures, semaphores, and so on.

During my ECE & CS studies, aside from my student job, I got exposed to:
  BASIC (1 day) and FORTRAN (both while still in HS), ALGOL (2
dialects), at least 4 different assemblers/machine languages, SNOBOL,
LISP, PLUS (Programming Language for UNIVAC Systems - their PL/S),
AMINOL (You won't find anything about that anywhere any more), PLUM
(Maryland's PL/I subset).

During my student jobs I picked up 1410 Autocoder, COBOL and a tiny bit
of System/370 Assembler.

I learned about data structures up the wazoo: linked lists, stacks,
activation records for block structure languages (3 times), and heaven
only knows what else, with exposure in many courses.  Also
multitasking/multiprocessing and all that went with that in multiple
courses.

And performance measurement and management.

[FWIW, this was a time before PASCAL and databases]

I got exposed to soooo much stuff that picking up anything else later
was essentially trivial.  So, when I got to a 360/65-MP and had to do
Assember and COBOL no biggy.  When we had an Amdahl 470 that was fast
enough that poorly designed software had threading issues (as in an
idiot developer of a commercial product that expected to use MVS task
dispatching priority to avoid deadlocks), that was easy too.  As was
teaching my co-workers how to use a stack structure to implement a
recursive algorithm without having to do recursive subroutine calls.

I was able to explain to the other system programmers what happened when
utilization of a resource gets to me more than about 80% (think disk
queue) and how fast things go to Hades in a handbasket.  Queing theory
came in handier than heck when I was able to model the effect of
response time store and retrieve image documents and demonstrate to our
management and our vendor that we would have lines around the block -
the project got delayed a year as a result.

That is the kind of thing a Computer Science program ought to be doing.
 Teaching advanced techniques and concepts that can be used in
situations folks haven't even discovered that show up 30 years hence.
These are things that MOST ordinary programmers are not likely to learn
or even need as long as they have access to that kind of expertise.
Some call it "ivory tower", but I hope my explanation has shown that
there is MUCH more to it than that.

(Of course, there are some who learn such stuff and can never apply it,
just as there are engineers that can't apply knowledge in that domain,
either.  That isn't the fault of the university.  That is the fault of
the student or whomever pushed the student into a situation they were
either not a good fit for, or were not sufficiently prepared for.)


BUT, outside of a 4 year degree environment, ah, that is an entirely
different story.  THAT is where we should be teaching "real world
problems".  Stuff that folks can use immediately to land, keep and
succeed at a job.  How to do the things that developers encounter each
and every day.  Current technology and practices and languages that
folks are actually using every day.

Think about it as an analog between carpentry and structural
engineering.  When a carpenter really needs engineering help, they go to
an engineer (for example, as the contractor did to cantilever the deck
off the back of my house), but that doesn't mean that every carpenter
should have all the complex skills an engineer has, nor does it mean
that an engineer needs all of the skills a carpenter has - though of
course the engineer had better know practical and doable from crazy.  ;)

Trouble arises when you try to teach EITHER ONE broadly/completely in
the other environment.  The (MIX) assembler course I took focused WAY
too much on "niceties" and trying to squeeze out every last bit of
performance, instead of teaching conceptually how to wring bits from a
turnip.  It was a horrible waste of time (and pretty much out of date
before it even started.)

There can be exceptions, of course. There is a place, for example, for
teaching non CS majors in a university setting Python (not my favorite)
because they may need it in their studies and research these days.
Sorta like teaching Calculus to broad numbers of student who are not
math majors.

I tried to teach my Linux OS students at a tech school a little bit
about how virtual memory worked so that when faced with real world
performance issues they might better cope rather than just rote numbers
- with mixed results at best, but hopefully some of it stuck.  Same
thing with multi-processing and what could go wrong with programs and
systems in that Linux OS class.  But I would never ever dream of
teaching them LISP -- wrong place for that.

</RANT>

JRJ


More information about the cctalk mailing list