In article <9409191742.AA15343@ah.com>, Eric Hughes <hughes@ah.com> wrote:
I'd suggest that a much more productive avenue of approach would be to improve the aliasing facilities of a remailer provider to allow a pseudonym to look like a fully normal name.
I'm not sure that's a good solution.
Todd, Todd, Todd. You can run a remailer and the mailing list on the _same_ machine and do the aliasing in the remailer. You can even restrict operation of the remailer to work only with the mailing list, if that's what you want.
The issue here is clean separation of abstraction.
Well *excuse me* for being clinically thick... I shouldn't post after more than 20 hours w/out sleep. You're right, of course. Though the remailer and the mailing list software would probably require some hacking to make the coupling tighter, in the process giving both limited-use remailers (probably undesirable in the generic case, but I can think of special uses) and access-controlled mailing list software (definitely uses for this, as some exist).
This doesn't require AMS. I've done the same hack myself in ruleset 0 of sendmail. Then you tweak the HReceived line to add the $u macro, which under sendmail v8 includes the whole address which caused delivery.
Could you send me what you've done on this? I think it's a desirable feature to have, though requiring that people hack their sendmail.cfs is not a big boost to the "popularity of package" indicator. -- L. Todd Masco | "A man would simply have to be as mad as a hatter, to try and cactus@bb.com | change the world with a plastic platter." - Todd Rundgren