Re: Garbage Collection - The Trash Begins To Pile Up
Kevin Hall wrote:
I agree. But if the burden is more on the OS people, how can you blame
the C++ standards comittee for not recognizing the problems? I bet
that if you asked any of the C++ comittee members, they would
acknoledge these problems and as you mentioned say there needs to be
more standardization on the OS side.
I find fault with the OS people for not finding a regular model for
synchronization. I find fault with the C++ Community for presuming
that the lack of regularity can be fixed by tweaking the language
somehow. If the C++ Community will acknowledge that the OS folks need
complete their frameworks first, then we are in agreement.
Note that I said "fix", not standardize. It would find it entirely
acceptable for example, if, on each OS, the primitives were
present-but-ugly. C++ could then clean up the API's easily.
What I would caution against is expecting the "fix" itself to be
provided by C++. That, IMO, is not possible. It requires kernel
rework, and the ratio of language-improvement to OS-improvement is so
low, it would probably be more productive to reallocate the brain
cycles asking "What could be done to make C++ behave like kernel-mode
primitives?" and instead apply those cycles to asking, "What is the
regular set of user-mode/kernel-model primitives that will provide
closure in the synchronization space?"
Someone mentioned recently that system-wide named objects have
relatively recently found their way into Linux. This, IMO, is a great
example where someone, somewhere, thought long and hard and realized
that you probably need that.
-Le Chaud Lapin-
--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]