[Info-vax] PERL for OpenVMS Alpha that will install on V8.4

[Jan-Erik Soderholm]

David Froble wrote 2013-12-03 06:33:
> Stephen Hoffman wrote:
>> On 2013-12-02 09:20:01 +0000, Jan-Erik Soderholm said:
>>
>>> The 10 line example on Hoffs link runs out-of-the-box with the Python
>>> port for OpenVMS.
>>
>> This also ties back to my earlier comments about how little source code
>> can be needed these days to produce results that would have been
>> impressive achievements in an earlier era and with older tools; of how
>> much more dense and effective newer code and newer tools and libraries
>> and frameworks can be.   The older compilers and older environments are
>> feeling rather like coding in assembler.  Which is part of why you want
>> to learn newer tools and newer environments, too.
>>
>
> Actually, there is most times a large amount of code behind such
> simplicity.  The library procedures are of course rather nice.
>
> But who is to say that "legacy" (I hate that word) languages cannot have
> such libraries also?  Consolidated Data has a huge number of library
> routines to support Codis, and therefore, implementing new functions in
> Codis can be rather trivial.
>
> If anyone is going to do comparisons, at least try to keep it apples vs
> apples and oranges vs oranges.
>
> I'm going to raise the question of whether the libraries in some of the
> "newer" environments might be more bloated that similar capabilities
> implemented in "older" languages?
>

Yes they are. Running a Python script from DCL includes a 1-2 CPU sec
startup/init of the Python environment. And Python itself is of course
a bit more "CPU-hungry" then the equivalent Cobol code would have been.

Also note that a Python module can be written in C, and as such has
more or less the same performance as if called from C, Cobol or whatever.

I use Python for CGI-scripts called from the WASD webserver. In that
case there is a special Python environment that is persistent and that
loads the .py files only once and saves them in internal format. First
call from a browser takes 2-3 CPU sec and the following are realy fast
only depending on the actual work to be done. In the 0.01 - 0.10
range for simple tasks (a fetch from Rdb and HTML formatting).

> "New" doesn't impress me ....

Few things impress me, but I think I can select the right
tool for the job at hand.

Jan-Erik.