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

[Arne Vajhøj]

On 11/5/2023 9:18 PM, bill wrote:
> On 11/5/2023 9:09 PM, Arne Vajhøj wrote:
>> On 11/5/2023 8:41 PM, bill wrote:
>>> On 11/5/2023 7:39 PM, Arne Vajhøj wrote:
>>>> 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.
>>
>> Neither.
>>
>> But Python code may (and for scientific computing usually do) use
>> libraries written in Fortran.
> 
> Oh...  If all your talking about is libraries there really isn't any
> reason I can think of why any language can't call libraries written in
> a different language.  I know that I can certainly (and frequently do)
> call C libraries from COBOL.  And back in the good old days we used to
> call the math functions from NAGLIB in lots of different languages.

It is quite common.

The interesting aspect here was not the use of a library
in another language, but that the "new" language in the domain
Python under the hood use the "old" language of the domain
Fortran.

> But I have to ask the question...  If Python is so good why are the
> libraries written in Fortran?

Different languages are good for different things.

Python may only require 1/10'th of the lines of code
that Fortran does to express where to load matrices
from and how to multiply and invert them, but Fortran
may be a factor 100 faster actually doing the matrix
multiplications and inversions, because it is compiled
unlike Python that (commonly) is interpreted.

So expressing what you want done in Python and
using Fortran libraries combine the best of both
languages.

Arne