From: Kai Henningsen (kaih@khms.westfalen.de)
Date: Sun Jul 01 2001 - 06:25:00 CDT
peteralf@fabel.dk (Peter Alfredsen) wrote on 22.05.01 in <734f1da984abc7dded6a11044dec498f@peteralf.sewer.output>:
> On Mon, 21 May 2001 14:53:39 GMT, chl@clw.cs.man.ac.uk (Charles Lindsey)
> wrote:
>
> >Yes, I agree we need some text. but it is rather easy to pick holes in
> >that one :-(. Essentially, we need to make the basic point, but without
> >getting tied up in too much detail. The more detail you provide, the more
> >holes you can pick :-) .
>
> Yeah, okay that text didn't really bring across what I was trying to
> convey.
>
> The point I am trying to make, is that an archive should only be defined
> as an archive in the sense trying to be adressed with the Archive header,
> IF it is somehow served to the reader as a cross-indexing system, that is
> beyond that normally used in the reading of USENET. "Normally" is here
> used very vaguely because times may change and so may access methods.
May I point out that I consider anything without the ability to search
over all messages as unfit for use, and that my current newsreader as well
as the one I'm currently (very, very slowly) writing both have (or will
have) that feature (and do, in fact, both do it by carrying a separate
article copy)?
> Basically, I think that it is better to differentiate on the vast indexing
> that servers such DejaGoogle tend to do, which brings a persons entire
> usenet-history up on the screen in one single search.
As long as I still have the articles, it takes me maybe half a minute at
most to search through approximately 300K articles (1 GB) for those of a
specific poster. And I do in fact use this regularly (as in, at least
daily).
> If the server indexes by group, as is normally used, and there is NO
> crossgroup indexing, only indefinite storage, it is a USENET server, not
> an archival agent.
It's also, IMAO, rather crippled.
MfG Kai