Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!usc!samsung!uunet!microsoft!steveha From: steveha@microsoft.UUCP (Steve Hastings) Newsgroups: comp.os.msdos.misc Subject: Re: Microsoft's XMS spec Message-ID: <55886@microsoft.UUCP> Date: 17 Jul 90 17:54:33 GMT References: <15173@reed.UUCP> <430@beartrk.beartrack.com> <10502@ists.ists.ca> Reply-To: steveha@microsoft.UUCP (Steve Hastings) Distribution: na Organization: Microsoft International Products Group Lines: 29 In article <10502@ists.ists.ca> pasquale@sgl.ists.ca (Pasquale Leone) writes: > > Many have complained that their program that can use XMS memory >won't work under mswindows 3.0's HIMEM.SYS . This is because windows 3.0 >and the new himem.sys does not comply with the XMS (eXtended Memory Spec) >as originally set out by microsoft and as used by MANY other companies. >(Desqview, Rational Systems, 4dos, etc) >They seem to think they can force everybody out of business by setting a >new incompatible "standard". This is a needless flame. The new HIMEM.SYS is, I believe, compatible with DPMI, the DOS Protected Mode Interface standard. Microsoft has worked with a number of other companies to hammer out a spec that will be open for everyone. (I don't have the list handy, but I remember that both Quarterdeck and Rational Systems were on the DPMI committee.) It's true that the DPMI standard is different from the old DOS extender standards (like VCPI). Microsoft took a lot of criticism for this. But as far as I can tell it was an honest desire to make better software (crashing less often) behind this. Gordon Letwin posted a long and detailed explanation of why VCPI had to go, and why DPMI is better and worthwhile; I wish I had kept a copy. I am not a systems guru and there may be another reason behind this change to HIMEM.SYS. It could even be a screw-up by someone at Microsoft. But I know it is not an attempt to force anyone out of business. -- Steve "I don't speak for Microsoft" Hastings ===^=== ::::: uunet!microsoft!steveha steveha@microsoft.uucp ` \\==|