[5903] in www-talk@info.cern.ch
Re: Forms support in clients
daemon@ATHENA.MIT.EDU (Darren New)
Thu Sep 29 00:18:19 1994
Date: Thu, 29 Sep 1994 05:15:32 +0100
Errors-To: listmaster@www0.cern.ch
Errors-To: listmaster@www0.cern.ch
Reply-To: dnew@sgf.fv.com
From: "Darren New" <dnew@sgf.fv.com>
To: Multiple recipients of list <www-talk@www0.cern.ch>
> Actually, we didn't include the notion of any intermediate relays, it
> was pretty much peer-peer. I certainly don't want to have to hack
> sendmail.cf to make my web stuff work.
Well, then you need to queue stuff locally at the server and retry
periodically, and hope that it isn't a SLIP connection that just doesn't
happen to be up when you're polling.
And sendmail != SMTP. Write something that translates sendmail.cf into
something legible for a new mailer and I expect people will flee sendmail.
> History, and because much of the world has differnt points of view
> on what e-mail should accomplish.
And that it's the only protocol that already has a bunch of functionality
that you don't get with IP's point-to-point realtime connections.
Now with MIME, you have typed data that can invoke arbitrary programs, either
when the mail arrives or when the user "reads" it.
> I was amused by a recent posting suggesting that to receive e-mail
> based asynchronous replies from a server my web viewer program should
> edit my /etc/alias file, run newaliases, and push my YP/NIS maps. I
> wonder what happed to "safe"?
That's many more levels of cruft than you really need. Why would you need to
change the routing of email to handle a message saying that a new paper has
been published?
