Path: utzoo!attcan!uunet!husc6!psuvax1!gondor.cs.psu.edu!flee
From: flee@gondor.cs.psu.edu (Felix Lee)
Newsgroups: news.admin
Subject: (Re: listservers) and reader-initiated sendme
Message-ID: <3679@psuvax1.cs.psu.edu>
Date: 6 Jul 88 03:14:37 GMT
References: <3335@s.cc.purdue.edu> <11457@steinmetz.ge.com> <270@octopus.UUCP> <280@octopus.UUCP>
Sender: news@psuvax1.cs.psu.edu
Reply-To: flee@gondor.cs.psu.edu (Felix Lee)
Organization: Penn State University Computer Science
Lines: 26

In <11457@steinmetz.ge.com> welty@steinmetz.UUCP (richard welty) writes:
> In <3335@s.cc.purdue.edu> rsk@s.cc.purdue.edu.UUCP (Rich Kulawiec) writes:
> >Clearly, however, this is a huge lose for the originating site, which
> >must send 100 copies of something rather than 1.
> It can be a huge lossage for links adjacent to the originating site
> as well.

If you use reader-initiated sendme, then a site only has to send an
article once to its neighbors, when asked.

You might tell your newsfeed "send me comp.*, but not comp.sources.*,
but I want an index for comp.sources.*, so I can ask you for specific
articles."  Your newsreader program would show you the index.  If the
article you're interested in isn't already at your site, you'd ask your
feed to send it to you.  If your feed doesn't have it, it would ask its
feed for it, and so on.

Once your site receives the article, it stays there until it's expired.
An archive site would be a site that never expires articles.

This would have less article traffic, but more ihave/sendme traffic.
And there would be a significant delay when asking for non-localized
articles.  It's probably most useful for leaf sites and sites at the
fringes of the net.
--
Felix Lee	*!psuvax1!gondor!flee