Filtering was intended to do some things that were not obvious to mailbox contents; you can tell me if procmail could do these things to messages saved in a mailbox.
Our belief was that somebody has a mailbox that collected mail about a subject (or series of subjects) and wants to publish them on the web so that other people can easily see them. In the real world, such a mailbox might have messages that the publisher doesn't want on the web, and other messages where the "Subject:" field does not accurately describe the contents.
Our intent with filtering was to allow some after-the-fact corrections based on these realities. Our filters were designed to have three parts, a subject match, a date match, and an action.
The subject match could match any subject, or provide a string to be equal to the subject, or provide a string that might be contained in the subject.
The date match could match any date, or have a date before a specified date, or have a date after a specified date.
The action would be to either discard the message (meaning that it would not have an HTML page generated for it), or to set the subject to a new subject.
For example, if there were a chain of messages where people were replying to previous messages but the subject changed, then a filter could be used to reset the subject to something more descriptive, after a specified date.
That would make reading the HTML pages more informative.
The filtering was expected to be before hypermail processing itself was performed, with a process that would empty the "real" mailbox and create a "temporary" mailbox for hypermail input.
We had looked at more capable filters, but we had some difficulty finding an intuitive way to build them through a GUI.
Our filtering was seen as a way of doing "fixups", not of routing messages to a particular mailbox or blocking spam.
David S.Cargo Received on Wed 28 Oct 1998 07:27:47 PM GMT
This archive was generated by hypermail 2.3.0 : Sat 13 Mar 2010 03:46:11 AM GMT GMT