LATEST CHATTY HEADER
Subscribe to Shacknews Mercury starting at $1/month!
Chrome Shack Community Guidelines Chatty Search
Scroll down to join the conversation.
New to Shacknews? Signup for a Free Account
Already have an account? Login Now
Subscribe to Shacknews Mercury starting at $1/month!
Chrome Shack Community Guidelines Chatty Search
Scroll down to join the conversation.
http://www.news.com/8301-13860_3-9911470-56.html
So much for Vista I guess. Although it will be another 2 years until they get a majority of the bugs worked out.
Anyone heard what is big about this version?
Thread Truncated. Click to see all 231 replies.
The problem (back in 1989, ouch) was that you could only automatically prove rather basic properties of your processes (or collections of processes). You can't automatically prove you won't deadlock, or gobble up all available memory, or execute within a certain time period. And all those things are vital for any desktop operating system.
The nice thing about separating processes into different heaps is that you can absolutely control how much memory each one has, and you can quickly and safely kill it off if it locks or goes mad. Plus you get locality of reference. If you have a single huge heap, over time, it will get fragmented. Each process will be accessing pages everywhere and your caching will go down the shitter.
And it's not just memory. All resources need careful control, like file descriptors, graphics contexts, whatever. If you try to automate resource deallocation you are going to get severe performance problems. This is why .net gets rather tied in knots with finalize and all that.
I like the objc/python/etc. system myself. Have a refcounting system so that stuff really does get freed up when it needs to be freed up, plus a separate GC to break cycles. And having separate heaps gives you safety and isolation, and gets you reference locality too.
The post has been reported. Thank you!
You must be logged in to post.
You must be logged in to post.