James Byrne writes:
>I believe that all modems that respond to the Hayes command set
>will recognize [evil string omitted (GS)] as a switch to local
>command mode instruction, but only after a predetermined time
>of inactivity.
 
Hayes modems look for the sequence:
 
<DELAY> <PLUS><PLUS><PLUS> <DELAY>
 
where the delays ensure that the modem does not false trigger
on data that happens to include three '+' characters in a
row.
 
Up to a certain date, all "Hayes Compatible" modems do the
same thing.  Unfortunately, Hayes went and *PATENTED* the
delay part of this protocol.  They then went around to all
modem manufacturers demanding royalties for the use of this
feature.  Around the same time they ran a series of weird
ads in Computerworld etc. which seemed to imply that users
of modems that had not licensed Hayes technology would
actually come to some physical harm (really bizarre stuff).
People who didn't want to pay royalties to Hayes were forced
to remove the "infringing" feature from their products.
 
I find Hayes behavior in this matter so annoying that I will
no longer consider the purchase of any product they make.  On
the other hand, I have no sympathy for anyone who uses a
modem that is so stupid as to believe that interpreting you-
know-what in a datastream without delays on either side this
way is acceptable.
 
Actually, I believe that typically what they look for is
something like <evil string><CR>, so you will only have a
problem with a record that *ends* in '+' characters (as the
.sig in question probably did).
 
G.