 
Thu, 11 Dec 1997 18:53:31 +0100
|
On Wed, 10 Dec 1997 23:51:44 -0500 Don Wiss <[log in to unmask]> said:
>There are now 10 million AOL subscribers and many MSN subscribers. It is
>very hard for us to understand why L-Soft hasn't written a patch ASAP to
>deal with these non conforming errors, since it is such a common
>problem.
L-Soft releases a new version of LISTSERV about once a year, so this is
the time frame in which one can introduce a change to the installed base.
AOL on the other hand can change their bounce system overnight. When this
issue came up a couple years ago, they said they would change their
bounce format so that it would work with LISTSERV. This change is tied to
the introduction of their new, un-slow mail system which you can rest
assured they want to move on with *very badly*. For reasons that I do not
know but can imagine, the introduction of this new mail system keeps
getting delayed such that the implementation date is invariably "in the
next couple months", or so it seems. Meanwhile, probes work fine with AOL
and with nearly all other non-conforming mail systems (as long as they
send the bounce back to LISTSERV and not to the poster). I have also
noted a trickle of AOL bounces in a format understandable by LISTSERV, so
I imagine that they are making some live tests.
As for MSN, when implementing heuristics for a specific non-standard
bounce format, it is important to have some kind of assurance that the
format will not change tomorrow and break everything, and that the format
is consistent (that is, that there aren't "weird cases" that you haven't
seen and which the kludge you are writing will interpret as failures when
they aren't). Otherwise you are just cutting some rope with which to hang
yourself later. It is difficult to obtain this kind of assurance from
online providers. A much better approach is to work with the people who
develop the mail product they are using and get them to change to the
standard bounce format, which benefits a lot more people than just MSN.
Eric
|
|
|
