Re: Microsoft & Ada

[Brian May <bam@snoopy.apana.org.au>]

>>>>> "stephane" == stephane richard <stephane.richard@verizon.net> writes:

    stephane> *** Loss of standards, like what they did to C++, Like
    stephane> Borland did to C++ too...they didn't totally destroy C++
    stephane> but in most cases, code written for Visual C++ tend to
    stephane> not compile as is on other C++ compilers.

What if, say Microsoft were to design and implement thick Win32
bindings for functions that are not standardized in Ada? Some
examples: Win32 GUI or serial IO.

Would this be considered a good thing or bad thing?

What if there were able to implement the bindings in such a way that
they don't have to supply any source code? What if the bindings can
only be used with Microsoft's compiler? (I guess it may still be
possible to write an open source library that implements the same
package specifications).

My personal thought is that code that uses such routines should be
isolated, so you can still compile the program without these functions
(perhaps using alternate code). That is, if portability is required.
In practise, I could imagine code being written that requires the new
functions without any thought as to making it portable.

I think there are a number of related issues here, so I am curious
what peoples thoughts are on the matter.
-- 
Brian May <bam@snoopy.apana.org.au>