[Info-vax] Programming languages on VMS

Arne Vajhøj arne at vajhoej.dk
Wed Jan 24 15:07:35 EST 2018


On 1/24/2018 11:16 AM, Jan-Erik Soderholm wrote:
> Den 2018-01-24 kl. 17:03, skrev Arne Vajhøj:
>> On 1/24/2018 10:51 AM, Jan-Erik Soderholm wrote:
>>> Den 2018-01-24 kl. 16:36, skrev DaveFroble:
>>>> I'll preface the question by admitting that I view much of the "new 
>>>> languages" as some people who just want to "re-invent the wheel".
>>>>
>>>> Do the "new languages" actually present more and better capabilities?
>>>
>>> Is a screwdriver better then a hammer? I would guess that you
>>> understand that the answer is that "it depends"...
>>
>> Some languages are better for some things than for other things.
>>
>> Languages are designed with certain usage in mind. So they are
>> often good for that usage and rather poor for other.
> 
> Isn't that just what I just wrote? That "it depends"? Or am I
> missing some point you are trying to make?

I was sort of agreeing with you - just elaborating a bit.

>>> In our case, Cobol is "best" for our core applications since it
>>> gives smooth and fast applications for our end-users.
>>
>> I find it hard to believe that Cobol is the only language that
>> could provide that.
> 
> Now, the context was *our* core applications. But yes, strictly
> technically, one could write the applications in a number of other
> (compiled) languages, but Cobol gives stable code with few options
> to "shoot yourself in the foot".

My guess would be that Pascal, Ada and PL/I would provide the same.

>                                    We would not be able to rewrite
> all our Cobol code in Java (or Python), the performance would go down
> the drain. We need stable response times to the factory equipment
> including database updates.

Unless you have really hard real time requirements or are memory
constrained then a JIT compiled and GC language like Java (or C# if
it was a platform supporting that) should provide fine
performance.

Python would probably both raise safety issues and
performance issues due to its very dynamic nature.

> Now, just speculating, but I think that it is easier to read 20-30
> year old and badly commented/documented Cobol code than 20-30 year
> old and badly commented/documented C code. I might be wrong... :-)

That would be my expectation as well.

:-)

Arne




More information about the Info-vax mailing list