On Fri, 6 Jan 1995, Steve Hirby wrote:
 
>      Re: 3, Seems to me that there is a legitimate question whether LISTSERV,
> in implementing auto-delete, should knowingly rely on information that is
> unreliable.
 
It is not the fault of LISTSERV that some systems send erroneous error
messages; it is assumed the error messages are reliable.  I doubt there is
any other practical way to do it.  How would you have LISTSERV determine
which errors are "unreliable?"
 
>           The intention of auto-delete is to save listowners work; deleting
> in error based on reliable information has the opposite effect.
 
What?  If the information is reliable how can the deletion be in error?
I have *never* seen an autodelete that was in error.  The error message
fit the parameters, so the subscriber was deleted.  If the error messaage
was erroneous, that's the problem of the particular node, not LISTSERV.
The subscriber should complain to his local system people.
 
        Douglas Winship    Austin, Texas    [log in to unmask]
                    Secondary AUTOCAT Listowner
                              MEDLAB-L