WH>LA>    386HPFS was developed solely by IBM, after the breakup, Check it
WH>LA>    out with a hex editor, you will see that the only copyright is by
WH>LA>    IBM, whereas a lot of the Win/OS2 files still have joint
WH>LA>    copyrights.

WH>Another urban rumor bites the dust, I guess.  Makes the pricing
WH>structure even stranger - but fortunately we're upgrading from Warp
WH>Server Advanced.  In a loaded server environment it's worth the price
WH>even at list.

GW> I don't think Lee has it quite right. I remember Mary Moynaham
  > (she worked on OS/2 file system drivers while OS/2 was still at
  > Boca) saying that HPFS386 was written by MS in assembler, so there was
  > very little they could do with it. She also said they were considering
  > re-writing it in C to enable them to enhance it, and I suspect if LEE is
  > looking at a recent version this is why there is no MS copyright on it.
  > If I can dig out archives from the period, (if I still have them, it was
  > several years ago) I'll check my recollections and post any corrections.

   George, what I posted was worked out in discussions with Dennis 
   Tonn about 9 months ago, and from IBM documentation, including an 
   interview with the then OS/2 chief architect (whose name I can't 
   recall).

   I suspect what you will come up with is that HPFS was written by 
   MS, and the 386 extensions were started at MS and rewritten 
   entirely at IBM.

   If I get the chance, (and still have em), I'll post you some of 
   the messages from that thread...

   In any case, even if it was a re-write job, the effect is still 
   the same: MS has no rights to 386, which explains why they don't 
   support it. It also probably explains why it works so well.

   [Later...]

   Here's what Dennis had to say (7/98) about the topic:


====================[start schnip]============================

 HPFS386 was originally written for Lan Manager and IBM Lan Server 
(both OS/2 server products). There was both IBM and MS code (and 
copyrights) in the original. 
 I am not privy to the legal arrangements between IBM and MS, but my 
understanding is that IBM got paid more for MS use of HPFS386 than MS
got paid for IBM's use of this code. In fact, this was true for the
licensing ratio between IBM and MS during Win3.1 and OS/2 2.0 too (and 
probably one of the big drivers to stop Win3.1 support after Win95 was
released).

 The HPFS386 agreement was separated from the general discussions 
during the "split" between IBM and MS. Different rules applied than 
for the rest of the OS/2 product. 

 HPFS386 has been 100% IBM written for a while now (around Lan Server 
3-4 days they removed the last vestiges of MS code). It has never been 
100% MS owned (or written).

 Of particular significance is that NT has never supported the HPFS386 
extensions to the file system (integrated user access rights, etc). In
fact, Lan Manager versions have never supported the dynamic buffer
sizing and later extensions of HPFS386, so we can guess at the time 
frame when the MS code was phased out. 

 The "rewrite" you found discussed in the 92-93 time frame was an all 
IBM project. I doubt MS had access to the rewritten code (but anything
is possible). 


   Denis       

====================[end schnip]===============================


                                             LRA


 -- SPEED 2.01 #2720: What is another word for "Thesaurus"?
--- Platinum Xpress/Win/Wildcat5! v2.0