DD>> Please quote the Hayes definition of autobauding.  

PE> I don't have it.  Ask Bill for it, he is the one quoting it.

DD> And you believe him on this?  You don't believe his other statements.

Of course not.  I was being deeply sarcastic of him talking
through his arse about what is needed to be done to satisfy
find detail of the Hayes requirements that he's never seen.

PE> Bill's the one who says the USR conforms to the hayes specs
PE> religiously, which is why the ATZ problem exists.  

DD> Please quote Bill's quote then.

I did in a previous message.

PE> Now, I will tell you what *I* expect from auto-baud, and 
PE> that is to USE the speed that I sent the last AT command at.  
PE> That is what *I* want, and I want to know what software, or 
PE> what terminal you are using that you would possibly want ATZ 
PE> to not auto-baud, and instead use the speed in NVRAM.  Name 
PE> just one terminal or one comms program.  Then I will name 
PE> some terminals and comms programs that do NOT work that way.

DD> ALL terminals!  

Wrong.  A WYSE terminal, if set to 38400, and you type in ATZ,
does NOT expect the baud rate to change on it, and it to change
in concert with that.  It doesn't change speed at all.

DD> If your terminal sends an AT command to the Courier at a 
DD> different speed to that stored in its NVRAM, it will
"auto-baud" to that 
DD> speed.  You can dial out or do anything.  

Yes, most of the commands do work.

DD> If the AT command is ATZ - it 
DD> will ackowledge at the current speed, and then reset to the stored speed.  

Yes, which is not autobauding.  I want a modem that autobauds with
ATZ, seeing as every single terminal and comms program works that
way.

DD> As soon as your terminal sends its dial command at the speed that is now 
DD> different to the modem's, the modem will "autobaud" to
that speed and dial 
DD> out.

Yes, everything except ATZ autobauds fine, I know.  We established
that 2 weeks ago or something (Bill documented that, and I confirmed
it).

DD> The locked speed characteristic is only revealed when you reset the port to 
DD> a different speed to your software and wait for an incomming call.

It is revealed when the modem fails to autobaud on ATZ, yes.

DD> If one is going to use ATZ (reset) to initialise the modem in the situation 
DD> where the modem is answering the call, one needs to ensure that the 
DD> software is listening to the port at the same speed the port is being reset 
DD> to.

Yes, to work around a USR design flaw, I know.  You don't have to
do that with decent modems.

PE>> If you just want to say the USR doesn't autobaud 
PE>> properly/at all, then that is fine, fucked by design, no 
PE>> problems. 

DD>> The USR Courier "auto bauds" as well as any other modem - 
DD>> HOWEVER when you 

PE> No it doesn't.  It doesn't auto baud on ATZ.

DD> Your modem is set to 38,400 isn't it?  Change your software to 2400 and 
DD> send ATZ - do you get an OK?  It has "autobauded" to 2400
to acknowledge 
DD> your command HOWEVER your command has just told it to reset all settings to 
DD> those in NVRAM - one of these is the port speed.  

No I didn't.  I told it to reset all documented S registers and
documented commands to their last saved values, and AUTOBAUD to
get the speed.  Since I don't have a mythical VT770 terminal 
that expects the modem to change speed back to last AT&W, I want
the modem to AUTOBAUD, even on an ATZ.

DD> This is NOT a problem 
DD> unless you expect the modem to answer the phone.  

Which I do, so it is a problem.

DD> In this case you must 
DD> ensure that the stored speed is the same as the software speed (as stated 
DD> in the documentation).

The documentation does NOT say that it fails to autobaud on an
ATZ.  And even if it did, it means it is fucked by design.  The
workaround you have suggested above, which I was using 10
minutes after discovering this design flaw, does indeed provide
a WORKAROUND.

DD> Good - at least you are starting to read the manual and notice how the 
DD> Couriers settings and characteristics are different to Rockwell based 
DD> modems.

Yeah, in the area of autobauding, it is fucked by design.  BFN.  Paul.
@EOT:

---