Re: Ada GC

Robert A Duff (bobduff@world.std.com) wrote:
> You cannot say that of Ada or C or C++. However, that may change.
> The Intermetrics Ada compiler supports GC, and somebody is working on
> GC for GNAT. Perhaps someday people will be embarrassed to sell Ada
> compilers without GC support.


Perhaps only because of Java.


> That would be a good thing, so long as compilers continue to support
> the *option* of using by-hand deallocation instead, for those cases
> where GC is the wrong solution. It's best to let the people
> designing real-time systems make that decision, even if some of
> those designers might make the wrong decision sometimes.


> Some have pointed out that is difficult to *require* GC in a language
> standard, because it's difficult to *formally* specify what it means
> to have GC.


But this didn't prevent the Eiffel or Java designers from writing, in
the first papers, in effect, "GC is extremely desirable and the first
reference implementation is going to have it and you're an idiot to
not support it when feasible."


And Eiffel has GC and Java does too.


Yes, I think the fact that the same wasn't done in the Ada95 design
process, that it wasn't a big priority, was indeed "cowardice" as Dr
Baker said.


If you didn't have any customers who wanted it, it may be that anybody
who wanted GC wasn't even considering Ada to begin with and had no
contact with the community. Think of the customers who weren't yet
there.


> Henry is unfair, I think, to criticize the designers of Ada 83 for
> ignoring GC. In fact, it is clear from the Ada 83 Reference Manual,
> that it was intended that GC be supported -- in the same sense as in
> Eiffel -- an intent, but not a formal requirement. However, the
> compiler writers did what their customers wanted, which was not to
> support GC. (Again, Henry may think those customers were foolish, but
> that's beside the point.)


I think this is conclusive empirical evidence to not use the same
agenda
"just do what our existing customers say they want" all over again.


You don't usually get second chances.
--

Return to the comp.compilers page.
Search the comp.compilers archives again.