Subject: | |
From: | |
Reply To: | |
Date: | Fri, 12 May 1995 11:59:05 -0700 |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
Re: programattic vs. command-level "LISTF,ACCESS".
Both Tom and Guy forget two critical items:
1) the Contributed Software Library
2) the 'net
I'm sure that free programs to use the "intrinsic interface" would
pop up immediately...programs that could easily be installed in
your local shared "CMD" group (thereby providing apparent CI command).
Keep in mind:
1) a command interface is limited...you *can't* write reliable
programmatic interfaces from a command interface, by definition.
(nor efficient ones, either)
2) a programmatic interface allows a command interface to be added.
Given the choice, I'll *always* vote for a programmatic interface first
and a command interface second ... and I'll almost always want both!
Guy writes:
> CI level access serves the entire user base.
Actually, it doesn't. With programmatic interface, free software
could be written that would be backward compatible with older
releases (e.g., use programmatic on 5.1 or later, AIFs on 3.0 through
5.0). ... a backward compatibility precluded with a CI approach (yes,
I'm excluding AIF:PE).
> An intrinsic serves only those with time and compilers.
No...an intrinsic is an *ENABLER*...
it's like the saying about giving a man a fish and he eats for a day,
or a fishhook and he can eat forever.
The fishhook is an enabler...just like an intrinsic.
Keep in mind the dialog over the format of a command...we already know
that all requests won't be implemented. With an intrinsic, you'll be
able to pick & choose (or write!) your desired command formats.
SS
|
|
|