BG> I don't accept that the failure of ATZ to auto-baud the modem
BG> is a fault or a bug at all.  In fact, I'm more inclined to
BG> suspect that it's been done that way intentionally for some
BG> obscure reason (but yes, it should be documented).

PE> It's fucked by design if done intentionally.

BG> Normal Hayes convention has ATZ resetting ALL
BG> registers to their previously stored values,

Nope, a real Hayes doesnt do that with the port speed
and the absolute vast bulk of other modems dont either.

BG> and USR appears to follow this convention to the letter,

Nope, it appears to be the exception on that ATZ behaviour
with the speed the RING is sent out at when the port speed
used in the ATZ command doesnt match the port speed which
is stored in the NVRAM and used on a modem power cycle.

BG> which is why the modem won't auto-baud, but reverts to the last
BG> port setting stored with &W.  Nothing wrong with that at all.

Corse there is, its completely counter intuitive, and the vast
bulk of modems dont do it like that. There is a remote possibility
that USR alone know of a good reason to do the port speed in such
a quirky way, but if thats true it should be very unambiguously
documented in the manual because its not what modems normally do.

And I think the possibility that there is actually a good
reason to do it the way that USR does it is extremely remote
when the vast bulk of other modems dont do it that way, and
you dont see reports of problems because they dont.

BG> Can you give me just one valid reason for even
BG> needing to alter the port speed on the fly?

PE> What do you mean "on the fly"?  What happens is I decide
PE> to find out whether my system can cope with 115200, so
PE> I change my terminal speed to 115200 in my comms program.

BG> That's what I mean by "on the fly".

Its not really a valid description. The reality is that most people do
go to some trouble to configure their modem the way they want it and use
ATZ in their modem init string. It may be MANY MONTHS LATER that you try
a port speed change for some reason or just use the modem on a different
machine or comms prog and have that fang you if you dont remember that
you have to &W at the different speed to get the RING to be sent at the
speed you are using. Fucked, completely fucked.

Like Dave Begley said, things like modems should be designed with a
minimum of quirky little gotchas like that. PARTICULARLY when you have
an extremely clever design which allows the modem to auto detect the
current port speed in use, and you basically toss that capability in the
bin on an ATZ for no good reason at all. Totally fucked by design IMO.

PE> I don't expect to go and have to issue a command to the USR to
PE> save to NVRAM, that's what auto baud rate detect is all about.

BG> No, you try it first, and if it handles 115200
BG> just fine, THEN you write it to NVRAM with &W.

*IF* you remember that you have to. You shouldnt have to.

BG> If one occasional extra AT command annoys you,
BG> I think you're going to be awfully disappointed in life.

Fucked designs do. Particularly when they completely fuck up a
brilliant concept on auto detection of the port speed in use like that.
@EOT:

---