On Thu, 10 Oct 2002 13:11:20 -0500, Trey Nash <[EMAIL PROTECTED]> wrote:
>I'm glad that I, at least, found some validation of my concerns regarding >these issues. It's clear that there needs to be the deterministic >destruction idiom available for use as well. And, IMHO, the dispose >pattern is a clunky solution. I would say that based on the lengthy discussions about deterministic finalization that has occurred in the past in many different forums, your points have validity. However, given these two choices (which represents today's technology) 1. A mark and compact GC system that handles circular references, allocates memory faster than a speeding bullet and results in little code bloat but I have to manually dispose of non-memory resources. 2. A reference counting system that deterministicly releases resources but I have to worry about memory leaks due to "dead" objects that only hang around because they reference each other. I'll take the GC because remembering to do a Dispose() in my opinion is easier to deal with than the circular reference issue. However, it would be way cool if Chris or someone else figures out a way to marry these two schemes in a performant way. -- Keith �And you�re probably saying, Bucky, I change strangs all the time. Well, I got news for ya boy; you ain�t!�. (VBTV - Visual Bucky's Toothy Tip http://msdn.microsoft.com/vbtv episode1) You can read messages from the Advanced DOTNET archive, unsubscribe from Advanced DOTNET, or subscribe to other DevelopMentor lists at http://discuss.develop.com.
