[Info-vax] OpenVMS async I/O, fast vs. slow

[bill]

On 11/5/2023 7:39 PM, Arne Vajhøj wrote:
> On 11/5/2023 6:14 PM, bill wrote:
>> On 11/5/2023 4:50 PM, John Dallman wrote:
>>> Incidentally, the situation for Fortran is rather different from COBOL
>>> and PL/I. Academic computer scientists don't usually touch it, but
>>> physicists, computational chemists and the like still use it heavily. So
>>> there are still people coming onto the job market who know it.
>>
>> Yes, but how well (and I don't mean syntax)?  Who is teaching them?
>> If they are not getting this training from CS Departments are they
>> getting any of the fundamentals or just syntax?
> 
> Scientific computing is pretty different from business applications.

I never said it wasn't.

> 
> I don't think trying to apply best practices for business applications
> to scientific computing makes sense.
> 
> The business application will need to be maintained 10-20-30-40-50
> years with constantly changes.
> 
> Many scientific applications (not scientific libraries those are
> different and more like business applications) only need one
> successful run and then the code goes on the shelf for future
> copy paste.
> 
> Very different contexts.
> 
> Fortran was very widely used back in the days in physics, astronomy,
> chemistry, biology, medicine, economics etc..
> 
> Some probably still use Fortran.

And I agreed with all of it.  They were called Domain Specific
Languages.  Interestingly enough the same article in CACM I
mentioned earlier that said IBM intended PL/I to replace all
the other languages had a comment about not understanding why
there we even had COBOL, Fortran and ALGOL.

> 
> But a lot of that stuff are done in Python today. Today
> Python is the language for scientific computing.

Python is a general purpose language that everyone wants to use for
everything.

> 
> Note that the interpreted Python is of course only
> "orchestrating" the number crunching - the number crunching
> itself are done in native code - Fortran, Fortran converted to C
> (yes!) or C.

Well, when you come right down to it it's all done in machine
language.  :-)

> 
> So the Fortran code still exist, but the scientists don't use
> it and don't even see it.
> 

I have to admit I have never seen anything about Python generating or
being converted into Fortran.

>> Some of the worst business programs I ever had to work with were in
>> Fortran and written by engineering faculty who needed something to do
>> during the summer.
> 
> Writing a business application in a language designed for
> scientific computing by people with mostly scientific
> computing experience has a very low chance of success.
> 
> Writing a nuclear explosion simulation in Cobol by
> people with expertise in ISAM files and BCD would probably not
> go well either.

And that was my point.  Use Domain Specific Languages and use people
who understand what they are doing to write the program.  Not general
purpose languages written by people with no specific expertise at all.

bill