[Info-vax] For sale: VAXstation 4000/90 128MB Fully Working and Tested

[Arne Vajhøj]

On 7/2/2022 8:26 AM, Bill Gunshannon wrote:
> On 7/1/22 20:41, Arne Vajhøj wrote:
>> On 7/1/2022 6:45 PM, Bill Gunshannon wrote:
>>> On 7/1/22 16:38, Arne Vajhøj wrote:
>>>> On 7/1/2022 2:25 PM, Bill Gunshannon wrote:
>>>>> On 7/1/22 11:05, Arne Vajhøj wrote:
>>>>>> On 7/1/2022 8:14 AM, Dave Froble wrote:
>>>>>>> It is not so much the quantity of "new" that is occurring, as the 
>>>>>>> contention of some that we all must embrace the "new", regardless 
>>>>>>> of whether what exists is working well and is not broken.
>>>>>>
>>>>>> I don't think the claim is that you should embrace the new.
>>>>>
>>>>> Of course it is.  Just look at OOP.  COBOL users refused to accept it
>>>>> because it really offered nothing they needed to get the job done and
>>>>> added layers of unneeded complexity.  The result was a full force 
>>>>> attack
>>>>> against COBOL that continues to this day.
>>>>
>>>> I don't think anyone is telling developers to embrace OOP.
>>>
>>> Of course they did.  Everything taught at University (at least on
>>> our side of the pond) went to OOP.  People, like the COBOL community,
>>> who refused to accept it became outcasts.  University's stopped
>>> teaching (or even discussing COBOL except to denigrate it).  Even
>>> CIS courses where COBOL was still the best fit dropped it.  I was
>>> there.  I saw it.  I fought it.  I still do.  But, alas, to no avail.
>>>
>>>> Modern software development is very much multi-paradigm. Procedural,
>>>> OOP, generic, FP and possibly with a tiny sprinkle of AOP. Developers
>>>> pick the tools they consider best for the task at hand.
>>>
>>> What non-OOP language is mainstream in Universities today?
>>
>> I believe there are still places where they teach C. Either
>> for OS or for embedded.
> 
> Not so much teach as read in things like OS classes where they still
> use Tannenbaum's text books.
> 
>> But more important they are not teaching OO only languages. They
>> teach multi-paradigm languages that support OOP and other paradigms.
> 
> Primary teaching languages are still things like Java although Python
> is moving up the food chain.  And OOP is still the primary paradigm.
> Unless your side of the pond is very different.

I don't see that at either side of the pond.

Java today is multiparadigm.

 >> Java was born with procedural and OO support - got generic support
 >> in version 5 and FP support in version 8. No builtin AOP support,
 >> but AspectJ provide both static weaving and dynamic weaving (the
 >> latter via Spring DI).

If you took 100 Java applications either written or updated in
recent time (last 8 years), then I would expect to see:

100 using OOP
100 using generic
50 using FP
25 using procedural
10 using AOP

>> Universities are not really pushing OOP. They are pushing
>> multi-paradigm.
>>
>> And if comparing University usage to industry usage, then
>> I believe they are over-prioritizing FP and under-prioritizing OOP.
> 
> If industry is prioritizing OOP it is only because universities pushed
> them that way.  Much like Unix and VMS.  VMS was pushed out of academia
> and Unix rose to the place of honor.  And we all know ow that stands
> today.

No. Universities push multi-paradigm with a slight overweight on FP.

>>>> I know you think the reason why Cobol is not in demand is that
>>>> universities does not teach it and attack it. But demand rules.
>>>> If the companies wanted Cobol for the new application they
>>>> create, then it would be Cobol. But they don't.
>>>
>>> But they do.  Recent surveys have shown that not only is the number
>>> of lines of COBOL not decreasing it is increasing.  The only major
>>> COBOL IS I have seen go away went away not because of a desire to
>>> use another language but because the maintainer of the IS has been
>>> unable to find graduates who know or are willing to learn the
>>> language.  They don't know it because it isn't taught and they
>>> aren't willing to learn it because they had professors like some
>>> of ours who repeatedly told them even learning the language was
>>> detrimental to their futures.
>>
>> You can always get people if you are willing to pay enough money.
> 
> Businesses don't work that way.

Of course they do that is how a market economy works.

There is a supply and a demand - and price is determined from that.

If price goes up the supply increases and demand decreases.

Adam Smith's invisible hand.

>> "not being able to get people" really means "the stuff I am
>> doing generate less money than what other companies are doing
>> so the other companies get the people".
> 
> Not always.  When one has spent four years (and a lot of money) being
> indoctrinated into despising one particular option the result will be
> very obvious.

Money matters.

There are other languages that has never been popular at universities
but the industry has been able to find people anyway.

VB6, VBS and VB.NET are good examples. Universities never liked Basic.
Typical did not like MS either. Those languages was practically never
taught - and I am sure Dijkstra was quoted quite a bit. But the industry
was able to hire millions of developers for those technologies anyway.
Students learned C++/Delphi/Java/C# and they got a good job offer
doing one of those and they accepted.

>>>>>> More like continuously evaluating whether new stuff has some
>>>>>> advantages over old stuff.
>>>>>
>>>>> Most of it does not.  It's the old risk/benefit argument.  Most of the
>>>>> changes foisted onto the IT world offered little if any needed benefit
>>>>> and brought a lot of risk that adversely affects business daily.
>>>>
>>>> If you look at the world, then I think you will see that companies
>>>> that are investing in new technologies thrive, while those that stick
>>>> to what just works fine as always dwindle.
>>>
>>> Most of the Fortune 500 still use mainframes and COBOL. Most major banks
>>> still use mainframes and COBOL.  Credit Card companies.  Airlines.  All
>>> of the major automobile companies.  All of the major Aircraft companies.
>>> The Government at all levels except maybe local who never made it past
>>> the PC.
>>
>> There are a lot of Cobol code running. And there will continue to
>> be for decades.
>> But the new stuff are done with other languages.
> 
> Recent surveys say otherwise.

Really.

Usually Cobol surveys prefer to discuss the size of the code base
not how much is being added.

>                              When you look at percentages ot languages
> used COBOL seems to be on the decline.  But if you classify work by its
> actual value, well, Candy Crush wasn't written in COBOL and neither was
> Minecraft.  But when it comes to banking, insurance and other really
> needed application....

Cobol is becoming a pretty small part of what the financial sector use.

They prefer C++, Java, Python etc. for new projects.

>> And companies slowly migrate off. Not many, maybe just a few percent
>> per year. But it accumulate over many years.
> 
> Very few of the important ones are migrating off and many of those
> that do migrate into failures.

Some succeed.

And it accumulates.

> And, once again, this is something I would have thought the VMS
> community was very familiar with.

Hmmm.

Question: how many VMS systems was around in 1992 and how many VMS
systems are around today?

Seems like a lot of companies managed to migrate.

Arne