Carey Bloodworth wrote in a message to David Bowerman:
DB>You stated that USR's implementation of v.32terbo was "proprietary
DB>and unable to work with other v.32terbo modems".
 CB> I didn't say that the 'official' v.32terbo speeds were not the same
 CB> and that USRs wouldn't connect at that speed.  Only that USR's
 CB> extension was definetly propreitary and wouldn't connect with other
 CB> v.32terbo modems. After all, both modems were labeled as v.32terbo,
 CB> but if a user tried to connect the two, you definetly would not be
 CB> able to connect at 21k, even though they both were obviously
 CB> v.32terbo modems and 'should' be able to connect because they both
 CB> obviously implemented the same protocol (after all, it _was_ same
 CB> name.)
Carey, do I have to quote that paragraph back again?  The section in quotes 
above is exactly what you said -- "proprietary and uanble to work with other 
v.32terbo modems".  Despite your claim, a non-USR v.32terbo capable modem 
would be able to connect at 16800/19200.  By your logic, a modem that was 
capable of connecting at 31200/33600 v.34 speeds prior to the release of the 
annex to the v.34 recommendation would qualify as proprietary and wouldn't 
connect with other v.34 modems.
Perhaps if you had said that other v.32terbo implementations only worked with 
USR's implementation at the AT&T specified bit rates and could not take 
advantage of the proprietary USR extension instead of what you did write, 
this discussion would not be necessary.
DB>to work with other v.32terbo modems".  As for the rest of your
DB>comments, could you supply a source for for your statement that 
DB>v.32terbo was public domain.
 CB> Craig Ford's own comm primer.  As he generally takes pains to make
 CB> sure it's accurate....
 CB> V.32terbo - 19200bps, with fall back to 16800bps. Designed by AT&T,
 CB> and is public domain, so any manufacturer can use this standard and
 CB> put it into their modems. USR has further extended this to support
 CB> a proprietary link rate of 21600bps.
"Was".  Not is.  At the time, as far as I am aware, AT&T charged for the 
license to use v.32terbo.
 CB> Note that he also calls USR's 21k 'v.32terbo' "proprietary"
 CB> extension. 
Yes.  That extension was proprietary to USR.  This did not stop another brand 
of modem connecting at the 16800/19200 speeds defined by AT&T.  Much like 
ZyXEL's extension to v.32bis to allow 16,800 bps connects did not prevent 
non-ZyXEL modems from connecting at v.32bis speeds to a ZyXEL modem.
DB>Given that USR's version of v.32terbo would connect with any other
DB>manufacturer's version of v.32terbo as evidenced by the number of
DB>people who connected to my system at those speeds, it's hard to see 
DB>where you found a basis for your claim that USR's implementation of 
DB>v.32terbo was unable to wo with other v.32terbo modems.
 CB> Not 'v.32terbo' itself.  Yes, that standard protocol worked among
 CB> modems.  It's the extension that USR made that wasn't compatabile
 CB> with other modems.  Even though it had the same name, it wasn't
 CB> capable of connecting at USR's 21k to other modems.
Carey, take the time to read the manuals.  USR referred to their extension to 
v.32terbo as part of their ASL package.  The manual made it clear that speeds 
up to 19,200 were standard.
DB> CB> All USR did was to create an incompatible extension with the same
DB> CB> name, and to confuse the protocol naming situation.
DB>What standard?  V.32terbo was an AT&T proprietary protocol which
DB>was never a ITU-T standard though AT&T did try to fit the name to 
DB>the ITU-T naming convention (V.27ter for example).  USR fully 
DB>supported the AT&T protocol wit
 CB> They named it that way because it was an offshoot of v.32.  They
 CB> realized that their extensions were different from the standard
 CB> v.32, even though v.32terbo could still connect at those rates, and
 CB> so they gave it a different name.
V.32terbo was not compatible with v.32/v.32bis.  A modem implementing only 
v.32terbo modulation would be unable to connect to a v.32/v.32bis modem.
DB>an extension that only affected v.32terbo connects between USR 
DB>Couriers -- that extension was covered under the ASL catch-all.  You 
DB>could always set S34.1=1 to disable the 21600 link speed.
 CB> Right.  But the points were
 CB> 1) The part of the message I originally responded to seemed to
 CB> strongly imply that USR created v.32terbo.  (As the discussion was
 CB> USR vs. Rockwell.)  In USR's own docs, no they don't say they
 CB> created it.  But the advertisements back then, the reviews, the
 CB> ads, etc. generally phrased things such as if USRs and everybody
 CB> elses was identical. I've seen other modem & bbs primers come right
 CB> out and say that USR created the v.32terbo standard, too.  Not just
 CB> the 21k extension, but v.32terbo itself.
Is that USR's fault?  Were they responsible for others' errors?  As for 
identical, any other v.32terbo implementation could connect to a USR 
v.32terbo implementation at all v.32terbo supported speeds.  How much more 
compatibility do you need?
 CB> 2) USR didn't create v.32terbo.  AT&T did.  AT&T was not mentioned
 CB> in that message.
 CB> 3) The only thing USR did with v.32terbo (beyond implementing the
 CB> AT&T version), was to extend it and still call the 21k connect
 CB> speed 'v.32terbo', which it wasn't.  The 21k connect speed was not
 CB> mentioned or covered in the 'official' AT&T spec.  Therefor the 21k
 CB> connect speed is _not_ v.32terbo.  It is v.32terbo 'like' etc.
Try v.32terbo with ASL or USR-V.32terbo.  As the Courier manual said, "at all 
standard rates up to 19,200 bps or 19.2K bits per second (bps)".  Given that 
the modem itself reports all v.32/v.32bis/v.32terbo connects as V32 in the 
modem response string, your argument is a trifle specious.  As for other 
differences, a USR v.32terbo modem could have assymetrical connect speeds 
with another Courier, again under the ASL catchall.  BTW, this also happened 
with v.32bis where a Courier to Courier connect could show a speed of 14,400 
in one direction and 12,000 in the reverse direction.
Regards,
       David
--- timEd/2 1.10+
---------------