Actually, it looks like the problem was the Answer String, not
the case-insensitive compare, because the compare looks like it
is harmless.  On the other hand, I tested with Brenton, and the
ATA vs ~~~~ATA makes all the difference.  I think the problem
is that on the first RING, Binkley sends ~~~~ATA, and that gives
time for the next RING to come in.  So ANOTHER ATA is probably
issued, and that is treated as a string to abort the first
connect.  :-(  Certainly Brenton can connect with ATA but can't
connect with ~~~~ATA as the only change made.  

However, there is another weird thing.  Brenton reports that the
tones my modem puts out are very weak with ~~~~ATA, but much
better with ATA.  That sounds preposterous to me.  Did anyone
else notice sound differences Saturday AM compared with Saturday
~midday and beyond?  BFN.  Paul.
@EOT:

---