PE> DD> If all modems use ATZ to reset the settings in memory to those stored
PE> DD> in NVRAM and if all modems store the port speed in NVRAM and if the
PE> DD> USR Courier is the only modem that resets the port speed on ATZ then
PE> DD> aren't all other modems fucked by design?

PE> FM> No. *If* all those things are true then all those other modems could be
PE> FM> said, in a legalistic sense, to have a bug in that their implementation
PE> FM> does not conform to the documentation.

PE> Even if all those things were true, which they aren't really,

FM> Oh, which aren't?

The first one, "All modems use ATZ to reset the settings in
memory to those stored in NVRAM".  As an example, my Netcomm
has stats on how many calls I've had.  Now do you reckon that
would be stored in NVRAM?  To help you make your decision, I
just did an ATI10 to look at the stats, powered off my machine,
done the command again, and all the stats are still there.

Now given that I do an ATZ after every call, do you think it
restores my stats to the time of last AT&W?

Of course not.  The objective is not to restore everything in
NVRAM blindly.  What it is designed to do is restore the
parameters you have saved with AT&W, e.g. if I set something
like S0=0 as documented in the manual, it makes sense to restore
that.  There is no requirement that you must restore everything
in NVRAM blindly.  You would have to be mad to think that modem
manufacturers aren't ALLOWED to keep undocumented information
for internal use only in the NVRAM of their modem, such as the
statistics AND the line speed.

And even if it were true, you would STILL need a quote that says
that the ATZ speed restoration overrides the AT autobauding.

PE> then you still have to show that the documentation and the
PE> implementation differ.  Let's see you find 2 quotes from the
PE> Spirit manual to that effect.

FM> Why do I need 2 quotes? 

One to say that it restores all parameters, including the speed,
the other to say that this command overrides the normal
autobauding of AT commands.

FM> p69 says re the ATZ command (inter alia) "Use
FM> the parameters stored in nonvolatile memory to set up operating
FM> characteristics." If we assume not unreasonably that
"parameters" means
FM> "all parameters" and "operating characteristics"
means "all operating

Well for starters I don't say that "parameters" means "all
parameters", as this is not a legal document or a standard
or anything, you can't expect them to cross the t's and dot
the i's.  I would interpret it as "the parameters that make
sense" to restore.  Do you really think you would have a leg
to stand on if you went to Digicom and said, "look here, it 
says parameters, not most parameters, and I consider that
the speed is a parameter too, fix your modem".  They'd laugh
at you, quite rightly too.  And even then, I consider 
parameters to be "ATS0=0" not "change the baud rate in my
comms program and remember to restore that as well".

FM> characteristics", and if I assume that the Spirit uses the the ATZ baud
FM> rate and not the one retrieved by the ATZ (I can't be bothered checking
FM> this, and I'm assuming from the comments from you and others in this
FM> thread that this is the case) then yes, the Spirit implementation does
FM> not conform to its documentation. IN A LEGALISTIC SENSE, as I said
FM> above. But it's still the right way to operate, it's just not documented
FM> that it does that.

And EVEN if you did use your definition, you STILL need a quote
that says that it overrides autobaud.

PE> FM> But it's the Courier that's fucked by design.

PE> Very true.

PE> DD> Isn't the Courier the only modem behaving as documented?

PE> FM> Yes, if all those things are true.

PE> It isn't even documented that it autobauds, nevermind which has
PE> priority, autobaud or restoring undocumented registers.  It's
PE> also undocumented that it is making sure it restores undocumented
PE> things at all, unless you make a few leaps of faith.

FM> Well obviously I haven't seen the USR documentation. I can only go on
FM> what's been discussed here, and in particular David's "if"
scenarios at
FM> the top.

Well it's the same leap of faith that you have above, that
"parameters" means "every single parameter, including things
that don't make sense to restore, such as stats, and things
that aren't actually parameters as such, just the baud rate
at the time of issuing the command".  In this case I think
they use the word "settings".

PE> FM> But if so it's the design that's
PE> FM> fucked, in that it does not conform to what users expect, and can
PE> FM> cause real operational problems.

PE> Yeah, regardless of the legalese smokescreen put up by USR bigots,
PE> the fact remains that the modem is fucked by design.

FM> I agree of course. But it does conform to it's documentation, assuming
FM> that says something like the Spirit does on "restoring operating
FM> characteristics".

The big killer is there is no mention of which is the overriding
factor, autobaud or ATZ.  The second killer is the leap of faith
in attempting to interpret words in the way you need to interpret
them to force it to restore internal-use settings.  BFN.  Paul.
@EOT:

---