Re: Basement tinkerers and inventors

On Sat, 1 Sep 2007 15:39:57 +1200, "Pete Dashwood" <dashwood@xxxxxxxxxxxxxxxxxxxxxxxxx>
wrote:

"Robert" <no@xxxxxx> wrote in message
news:q5lhd3phrd6b1sd4cqedl8c1qqf5e2s9t9@xxxxxxxxxx
On Sat, 1 Sep 2007 12:47:41 +1200, "Pete Dashwood"
<dashwood@xxxxxxxxxxxxxxxxxxxxxxxxx>
wrote:

The people who are REALLY into PC development are moving away from COBOL
because there are much better options available now for PC development.

Prgramming is programming. The operating system doesn't change the rules
of logic, Boolean
algebra nor good design. You are referring to a cultural change, not so
much a technical
change.

Did I say it was technical... :-)?

In fact, I see it as a complete paradigm shift involving technical,
cultural, and economic change.

In the Good Old Days, before 1990, the most important design constraint was inadequate
hardware. Compensating for slow hardware caused bad design practices to be accepted as the
norm. When machines became fast enough, some people didn't get it, they continued solving
a problem that no longer existed. OO was an attempt to break that by simply changing the
word order of programming languages. The old languages were verb initial, MOVE A TO B; OO
languages are object initial, B:COPY(A). Picture language in two dimensions, with verbs on
one axis and nouns on the other, grammar being the intersection of the two. OO rotates the
tableau ninety degrees.

Is that enough of a change? Is it a change at all? Some will quickly point out that the
word order of assignment, B = A, used in languages other than Cobol has always been OVS.
There is not a big difference between natural languages that are verb initial, for example
Gaelic and Biblical Hebrew, compared to most languages, which are subject initial. Oddly,
there are no object initial natural languages, which explains why Jedi sounds
other-worldly.

Experimental linguists have discovered that verb initial languages are the most difficult
to learn because of "information loss" during parsing. For more on this see
www.nbu.bg/cogs/events/2002/gruening.pdf

At another level, OO can be seen as a packaging convention that puts all logic related to
a data element (datum) in one program. In the Good Old Days, when we wanted to know
everything about an element, we had to search hundreds of source programs. Now we're told
it's all in one collection of methods. Is it really? The methods read "update element to
the value passed" and "compare element to the object passed". Now we don't search hundreds
of programs for the element name, instead we search hundreds of programs for the method
name. Is this an improvement?

If you're working daily in a world of Linux and Ruby or PHP or C# and
Windows, you really only have a passing interest in COBOL. (or none at
all...)

I worked in shop where we did all Cobol development and testing on
Windows, then deployed
on Linux without ever testing there. We could as well have deployed on
mainframe. Java
people are comfortable with the concept.

Is that shop still running?

ESRX is the darling of Wall Street, for having the greatest long term growth. If you had
invested $1K in ESRX in 1992, your stock would be worth $60K today. By comparison, the
same investment in MSFT would be worth $15K, in IBM would be worth $3K.

We processed and stored 50M transactions per month using an anemic Unix box that cost
under $10K.

Unix people
are as much into OO as PC people.

As far as I am concerned, Unix people ARE PC people. I use "Personal
Computer" pretty loosely. I klnow Unix is running on mainframes.

The computing landscape is no longer divided into two camps: mainframes and PCs (desktops
and laptops). On the contrary, those two are now on the sidelines. The main players are
'servers' running Big Unix and Big Windows. They are the core processors of most telephone
companies, airline reservations, credit cards, medical claims and other high volume
applications. Historically, 10K transactions per second was thought to be the limit for
the largest single computer. Now we have Unix boxes that routinely process 20-40K per
second. They do with 100-500 parallel processes, all doing the same thing. It's awesome to
watch, and even more of a rush knowing you wrote the code. :)

Because it wasn't invented in 1974, mainframers adopt a
wait and see attitude. They hope this too will pass.

They simply aren't PC developers.

They simply aren't Real Programmers.

Maybe.:-) The term "programmer" is taking on new meanings.

The word programmer is no longer used. We were engineers in the '90s, until real engineers
objected. Now we're developers. Those who couldn't make the cut are called testers and
production supporters.

.

Relevant Pages

  • Re: Forth as an operating system
    ... simplifies the Operating System services as it doesn't have to ... hostile, buggy, or badly written and protection is essential. ... That's not 'complete' compared to something like Unix. ... possible languages being hosted and being hostile. ...
    (comp.lang.forth)
  • Re: "Shanghai Stock Exchange" and OpenVMS
    ... The comp sci students wrote their own versions of things and added entirely new things; and then there was Berkeley Unix. ... Sometimes a task is more easily done on one of these than on VMS. ... It's not possible to support them all but I suspect that most of the major languages are supported. ...
    (comp.os.vms)
  • Re: Perl is a poorly understood language ?
    ... programmers as a scripting language suitable only to sys admins. ... them love to hate a certain set of languages. ... Another aspect of Perl is that it is very much within the Unix ...
    (comp.programming)
  • Re: "Have" as perfective auxilliary in various languages
    ... Germanic and Romance languages, for example, has occasionally extended ... means by "periphrastic resultative/perfect". ... point is to look at the result of the verb. ... German, amongst others. ...
    (sci.lang)
  • Re: Chez Watt Re: Common ancestor between man and ape
    ... If German word order is for English used, ... languages even or Sanskrit or the languages descended from it. ... There is one verb form for "It ... Some people proposed English prime which is English shorn of the verb ...
    (talk.origins)
Loading