Carey Bloodworth wrote in a message to David Bowerman:
 CB> I just saw this thread, and although I don't follow most of what
 CB> you were saying (or even care, really), I do want to comment on two
 CB> things you did say.
DB>arena?  As for V.FC, name any other manufacturer other than USR and the
DB>Rockwell chipset stuffers who produced modems that supported V.FC. 
DB>Now name
 CB> Hayes.  Rockwell & Hayes were the ones who developed V.FC
And which Hayes modems that did not use the Rockwell DSP supported V.FC?
DB>Now name the company that originated v.32terbo.
 CB> AT&T.  I don't remember what year, but it was definitly AT&T and
 CB> not USR or Hayes or Rockwell.  And they made their improvment
 CB> public domain, so anybody could implement it. USR made their own
 CB> modifications to it, making their version proprietary and unable to
 CB> work with other V.32terbo modems, but v.32terbo itself was done by
 CB> AT&T.
Really?  I had better tell those callers to my system who were using several 
brands of 19200 modems that they weren't connecting at 16800 or 19200 (AT&T, 
ATI and whomever implemented the softmodems).  The only proprietary part of 
the USR implementation that I am aware of was the extension of v.32terbo to 
21600. In many ways, a trivial extension much like the ZyXEL extension of 
v.32bis to 16800.  Due to the fragile nature of v.32terbo in general, an 
almost useless extension unless you had very clean lines -- the sort commonly 
found in test labs but rarely in real life.
 CB> Both bits of information comes from Craig Ford's excellent comm
 CB> primer that he posts in the Communications echo every month.
I seem to remember that what Craig placed in his comm primer was that USR 
extended v.32terbo to support a proprietary link rate of 21600.  I don't seem 
to remember any comments that USR's implementation would not work with any 
other v.32terbo implementation.
Regards,
       David
--- timEd/2 1.10+
---------------