On Thursday, July 13, 2000 2:28 AM, Geoff Bull
[SMTP:[log in to unmask]] wrote:
> > -- I tend to think that supersets are always a good idea
> I tend to think supersets are always a bad idea, especially
> when they are motivated by a desire to hijack standards
The POSIX Ada committee considered the issues with supersets. Without
supersets, one often cannot make full use (or sometimes even appropriate use)
of a system. With supersets, portability is severely compromised. Worse,
when a customer has two systems to deal with, and each has a superset, there
are inevitable conflicts.
The POSIX Ada committee considered this, and directed, within the standard an
appropriate method for providing supersets. The standard packages were
strictly non-supersets, though there were specific permissions for things
like adding comments. There were suggested namings and methods for adding
superset functionality. Those suggestions were designed with two purposes in
1. To identify, CLEARLY, when a POSIX superset function was being used
2. To separate diverse supersets in name space to prevent naming conflicts.
The problem with Microsoft supersets is that they are intertwined throughout
the language. Given some program written in Visual C++, it is virtually
impossible to "cut away" the superset usage and have any useful code left.
I believe that is what is meant by "hijack standards". I agree -- that is
reprehensible and damages real progress in software development.
"E"commerce is not life and death. If it was, it would be coded in Ada.
Steven Deller, Apex Ada Marketing
[log in to unmask], (410) 757 6924
Rational Software Corporation, http://www.rational.com
Apex email groups: http://www.rational.com/products/apex/usergroups
Apex technical notes: