| Richard 2004-11-17, 8:55 pm |
| Robert Wagner <spamblocker-robert@wagner.net> wrote
> I see little difference between the C standard saying 'the implementor
> defines the meaning of int' and the Cobol standard saying 'the
> implementor defines the meaning of extension types'.
Your point was that Cobol _provided_ machine dependent types, now you
agree that Cobol (the standard) merely didn't disallow them.
> At least Cobol provides portable types.
Exactly, all standard types were intended to be portable.
>
> The extensions are visible, reasonable and available to the user.
It is true that there are visible extensions in both MF and Realia.
However, there are invisible extensions in various MF compilers that
are not available to the user and are not documented.
Some of the x"nn" routines are documented in various versions of the
compiler, some such as x"91" only documents part of the functionality.
Others do _something_ but are 'invisible'.
If there were similar routines in Realia that were not visible then
you wouldn't know about them (becuase they are invisible).
> Realia was cleaner and documented. One wrote
> CALL 'MLI_XOR' USING A, B
> where MLI stood for Machine Language Interface. The CALL didn't call
> anything, it generated an in-line instruction.
Well _exactly_, your claim was that the compiler was Cobol, but it is
entirely possible that much of it is actually embedded assembler.
|