Re: flexfax: Q: Retrain Negative -- Giving up after 3 tries?

[David Woolley <david@djwhome.demon.co.uk>]

> 
> You could hand edit the info file to force 2400.  A reducing requirement on
> each retrain would presumably do the same thing.

Reducing speed on a failure is a legitimate thing to do for poor line
quality, but for a pure coding error it is useless, and for a coding
error introduced at source by a data underrun or race condition, any 
speed reduction should be permanent and independent of destination;
the fact that the problem is destination dependent would just indicate
a variation in the tolerance of the receiver to coding errors.  

Given the apparent modem dependence, it is possible that any underrun
or race condition may be in the modem.  In the case of an underrun,
particularly, it may be that some modems have less buffering than others.
In those cases there may be no reasonable work around, especially for
a multi-user platform, except for replacing the modem.

In particular, I'm concerned here that people are attacking the RTN symptom
when:

- they should be attacking the cause;

- RTN is being generated in error because the receiver can't distinguish
  between poor line quality and coding errors.