Path: utzoo!attcan!uunet!ginosko!xanth!ukma!gatech!hubcap!billwolf%hazel.cs.clemson.edu
From: billwolf%hazel.cs.clemson.edu@hubcap.clemson.edu (William Thomas Wolfe,2847,)
Newsgroups: comp.sw.components
Subject: Re: Pointers to code
Message-ID: <6341@hubcap.clemson.edu>
Date: 27 Aug 89 18:26:12 GMT
References: <6333@hubcap.clemson.edu>
Sender: news@hubcap.clemson.edu
Reply-To: billwolf%hazel.cs.clemson.edu@hubcap.clemson.edu
Lines: 21


   Randall described a few environments which provide "lightweight tasks";
   in such an environment, one can pretty much be unconcerned with the
   inability to directly invoke Ada subprograms by way of pointers, since
   one can simply enclose the desired subprogram in a task body; this gives
   complete flexibility, since one can store a pointer to a task and later
   invoke the task via the pointer, or even store the task itself as a
   record component and then invoke it directly.  This, of course, is
   completely portable, unlike the situation in which one insists upon
   keeping the code in the form of a subprogram and must therefore resort
   to machine language in order to get the subprogram invoked by way of 
   its 'ADDRESS (or by way of a pointer).  Assuming that lightweight tasking
   is provided (and Ada is largely based upon this assumption, certain
   obsolescent operating system technology notwithstanding), there is 
   really no particularly compelling reason to require the ability to
   invoke subprograms by way of a pointer, except perhaps as part of some 
   aesthetic desire for orthogonality.

    
   Bill Wolfe, wtwolfe@hubcap.clemson.edu