BG> &B1   Fixed rate.  Default.  The modem always communicates with the
BG>       terminal or computer at the rate at which you have set the terminal
BG>       or software, regardless of the connection rate. For the greatest
BG>       throughput, set the serial port to 115.2K, 57.6K, 38.4K bps for
BG>       high speed calls and to at least 9600 bps for 2400-bps calls.

BG>       This setting is not affected by the &N setting.  However, the
BG>       serial port rate must be equal to or higher than the &Nn rate.

RS> Note that this is the DEFAULT and that there is NOT A MENTION of
RS> &W at all. And when I dont believe that the AT standard actually
RS> does specify that the &W speed should be used in that case, rather
RS> than the speed of the last received AT command, it looks very much
RS> like a USR quirk instead. And an undocumented one at that.

BG> Rather than simply suggest that you don't believe the
BG> command is meant to work that way, it might have been
BG> better to provide some hard evidence either way.

Read the Hayes manual yourself, it doesnt specify that behaviour.

BG> Conjecture is becoming pretty pointless at this late stage.

It aint conjecture, we also have a number of modems which
clearly dont do it that way. Put that in with the fact that
even the Hayes modems dont, and you dont have any basis for
insisting thats the normal AT command behaviour on the RING text.

BG> &B2  Fixed for ARQ calls/Variable for non-ARQ calls. Answer mode only.

RS> Note this ISNT the mode he is using.

BG> Yeah, I realise that.

And thats the ONLY place the use of the port speed from NVRAM is
mentioned. It may well be that USR intended that approach with the
RING text speed, but they CERTAINLY never documented it at all.

RS> You could be forgiven for thinking that USR intends
RS> the comment about &W to apply to ONLY this config.

BG> A bit cryptic, I'll agree, but that's still not the impression that I get.

I think its fanciful to be suggesting that USR is attempting to
suggest that thats general behaviour with the RING text when its
only mentioned in ONE of the states discussing CONNECT behaviour.

BG> When the modem goes off hook and connects in ARQ mode, it
BG> shifts its serial port rate up to a user-specified rate, for
BG> example, 38.4K bps. If the connection is not under error control,
BG> the modem behaves as if it were set to &B0 and switches its
BG> serial port rate to match the connection rate of each call.

RS> AND this is clearly talking about AFTER the connect, NOT the RING.

BG> Pardon me for asking what may appear to be an obviously
BG> silly question, but under what circumstances would you
BG> expect the RING response to be seen at a different port
BG> speed to that locked by the modem and the terminal ?

The port speed stored by an &W isnt necessarily the locked speed at all,
its just what is used when there has been no AT command that can be used to
determine the port speed. Thats how all the other modems I have checked do it.

Clearly if you have never had an AT command at all since power on,
you have to use some speed, and it then makes considerable sense to
use the NVRAM speed, coz thats all there is. Makes no sense whatever
tho once you have receive a valid AT command and OKed it tho.

AND I havent seen any evidence that anybody except USR does it that way.

BG> If my modem's serial port AND term are both set to
BG> 57600 (which, I might add, they are), why would data
BG> arrive at the port at a speed other than 57600 anyway ?

If you have sent an AT command later at say 38400, AND OKed
it at 38400, its makes no sense whatever to be sending the
RING out at 57600, coz clearly thats not the port speed.

BG> To implement this feature, first set your software to the desired
BG> rate.  Then send the modem the AT &B2 [other settings] &W command.

RS> Again, it appears to be saying unambiguously that
RS> this only applys to &B2 and not the default &B1.

BG> Funny how intepretations can differ then, as I'd imagine otherwise.

Hard for you to substantiate tho when it only appears in the &B2 section.
PARTICULARLY when there is atleast one para of text talking about the port
speed detail in general, before it goes on to discuss each particular &Bn

BG> &B2 Fixed for ARQ calls/Variable for non-ARQ calls.  Answer
BG> mode only. When the modem goes off hook and connects in ARQ
BG> mode, it shifts its serial port rate up to a user-specified
BG> rate, for example, 38.4 bps. If the connection is not under error
BG> control, the modem behaves as if it were set to &B0 and switches
BG> its serial port rate to match the connection rate of each call.

RS> Soorree, either dud as designed, or dud documentation or both.

BG> What's wrong with that then?  It's perfectly clear to me what they mean.

Well, its an absolutely classic example of what can make some limited
sense when you know how it behaves, BUT its completely shithouse docs
in the sense that I bet no one reading that when first using a USR
modem would realise thats what they are saying, PARTICULARLY about
the speed the RING is sent at when its not even mentioned.

Let alone being shithouse coz it looks like that only applys to &B2.

And I cant see any good reason to do it like they do with the RING
anyway, particularly when most modems dont, and its quirky and very
counter intuitive, BUT if they insist one doing it that nutty way
ATLEAST they should point out that behaviour MUCH more unambiguously.

RS> Then there is the TINY matter that it makes no sense whatever
RS> to do it the way it does anyway, whatever they document. FAR
RS> too conterintuitive and no good reason to do it that way at all.

BG> That's bordering on gobbledegook, Rod.  :)

Nope, nothing remotely like it.
@EOT:

---