At 8:28 6/7/95, Jonathan M. Bresler wrote:
>at this time bulk_mailer does NOT probe MX records, but you have
>the source, this can be done ;)
At this point it was just an academic question <g>.
I am not yet a user of MD although in the next month or two I may end up
being responsible for a few lists. I also do not code Unix (When I DO code,
it is for IBM 370/390 Mainframes <g>). I am more in the position of being a
Systems Programmer saying to myself "What features or abilities would make
this product more efficient/useful" and thinking that not just
consolidating mail on the domain level but on the underlaying SMTP Server
level would make a program like Bulk-Mail more efficient (the pre-scan of
the DNS by Bulk-Mail would duplicate that which would be being done anyway
by the SMTP server and would just "prime" the local Cache so there would
seem to be little extra overhead to build an "alias" table and use it to do
the segmentation).
Also there might be some code to spin off smaller packets in the case where
there are a large number of addresses hung off of a single SMTP server (so
that no other domain/server is affected by an inability to establish
connect with that server as could occur when multiple servers are
associated with the same message [I am under the impression that SendMail
single threads the contacts so failure to deliver a message holds up later
addresses on the list - Which is the reason for Bulk-Mail in the first
place <g>]). Thus the spin threshold could be based not only on the number
of unique domains/servers but also on the number of actual addresses.
|
|
