Xref: utzoo comp.os.msdos.misc:2006 comp.windows.ms:12445 comp.os.os2.misc:1285
Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!usc!snorkelwacker.mit.edu!stanford.edu!agate!headcrash.Berkeley.EDU!magid
From: magid@headcrash.Berkeley.EDU (Paul Magid)
Newsgroups: comp.os.msdos.misc,comp.windows.ms,comp.os.os2.misc
Subject: Re: OS/2 2.0 is here! vs Windows 3.0 vs NeXT/MACH
Message-ID: <1991May9.024258.13117@agate.berkeley.edu>
Date: 9 May 91 02:42:58 GMT
Sender: root@agate.berkeley.edu (Charlie Root)
Organization: ucb
Lines: 18

In article <1991May9.003009.14478@netcom.COM> resnicks@netcom.COM (Steve Resnick) writes:
>There are other issues with Windows that aren't present in OS/2. In Windows,
>if you allocate memory, you are given a "handle" rather than a pointer.
>This handle must be locked and released as it is used. In OS/2 when
>you allocate memory, you are given a "handle" in the form of a selector.
>There is not need for the program to lock or unlock this handle becuase it
>is handled inherently by the procesor and the OS. Windows cannot do this 
>because it needs to be able to run in real-mode where you don't have selectors
>or a real virtual memory manager. I have several Windows programs which
>are direct ports to OS/2. These programs perform much like their Windows
>counterparts in that they deal with the locking and releasing of memory
>which ultimatly hampers performance.

Does this mean that if I get OS/2 I will never see the bug whereby windows
runs out of system resources. ie:  It has assigned all 8192 handles to
open windows.

Paul