Re: INFO-ADA Digest V93 #349

[pitt.edu!dsinc!gvls1!lonjers@gatech.edu (Jim Lonjers)]

Dave Emery said:
>>Much of the stuff you're asking for, particularly dealing with
>>directories, are not language issues, but operating system issues.
>>That's why there are C and Ada (and FORTRAN) bindings to POSIX.  

Mr. Keeler responded:
>Exactly!  But the C (and coincidentally C++) bindings are so well integrated 
>with the C language that many people do not realize when they are using o/s
>features vs language features, when they are programming in C.  Such is not th
e 
>case, and will not be the case (from what I heard at Tri-Ada92), when one wish
es
>to use Unix operating system features when programming in Ada.  It is my under
-
>standing that the Unix-C (C++) community has mandated that no POSIX bindings t
o
>any language other than C or C++ may be well integrated.  No doubt this
>unannounced "mandate" has been of considerable commercial benefit to the 
>C/C++ commercial community.  Of course one would expect that Unix, written in 
C,
>would have an advantage, but many of the additional raodblocks which have been
 
>put up, regarding whether thick or thin bindings may be included in the 
>standard, and what must be done to get it approved as a standard appear to be
>for other than techically sound reasons. (I have heard it suggested that Ada 
>should have its own operating system, but I think that this would serve to 
>isolate it still more.)  For Ada to be more widely used it needs strong 
>bindings to the operating system.  Without them, congressional mandate or no, 
>Ada is left out with the also rans.  

Just to clarify:  Ada POSIX is a standard.  It has been an IEEE standard
for more than a year and an ANSI standard for around 4 months.  Many regard
the binding to be reasonably well integrated into Ada, and in may ways,
superior to the C binding.  The Ada standard already handles Ada tasking
and the C binding will have to grapple with the serious problem of errno when
pthreads come into the picture.

Someone will have to educate me on what a "mandate" by the C/C++ community
regarding POSIX standards is.  There are a few misguided individuals who
believe that there is some sort of "language war" going on between Ada and
the C languages, and by somehow placing the Ada standards at a sub-par
footing with the C language standards, this will put C at an advantage in
the war.  Such an attitude is eschewed by most people I know in the POSIX
standards community.  The market decides how much money each language
receives--not standards.  By isolating POSIX into a C-only role only
dilutes the strength of the C standards.

I am, of course fully supportive of your view that Ada-only secondary
standards should be avoided as isolationist.  Such standards will only
serve to drive the cost of Ada development and deployment environments
higher.

Regards,
Jim Lonjers