I wrote:
> > In general, retries are increasingly delayed because
> > the retry cases include the case where the server
> > is overloaded.

Oleg Kobchenko wrote:
> I actually ran telnet against their mail server:
> there is no increasing delay.
 
The increasing delay I spoke of is the delay implemented
in SMTP software -- between automated retries.
 
If Yahoo is dropping connections, and SMTP software
is reconfigured to try ten times more often, that could
get interesting.
 
That said, perhaps what we're seeing here is a spam
avoidance heuristic -- spam software tends not to follow
the standards, and [for example] tends to not retry.  If
that's the case, I can easily imagine a future variant on
this system dealing with rapid retries in some different
fashion from the current system.
 
-- 
Raul
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to