I might suggest Resource Description Framework (RDF). I cann't speak for it, but it seems to be pushed by O'reilly and W3C.
http://www.xml.com/pub/a/2001/01/24/rdf.html http://www.w3.org/RDF/ http://ilrt.org/discovery/ Looks like they've got some open source supporting it on multiple platforms including java. The whole metadata area may not be completely ripe yet. There seems to be a lot of work being done still. I think I remember reading a paper about tring to distribute searchs through metadata. That's probably way outside the scope of freenet anyway. I like the idea of the user combining a few trusted indexies. I think google updates it's search data every 3 months or so. So I imagin making a fresh index every 3 months with maybe some patches in between. Question: how does the indexer find out about the sites to bulid an index? Was I right about nodes in being able sniff the public keys of freesits? Maybe the indexer could just request SSK@<public key>/<metaformatname>. If he cann't find it he could just assume the site didn't want to be indexed. I guess the other alternative would be sending a NIM to the the indexer. Chris --- michael <[EMAIL PROTECTED]> schrieb: > Toad <[EMAIL PROTECTED]> writes: > > > On Tue, Jun 24, 2003 at 11:27:41AM -0700, michael > wrote: > >> > >> So what we someone came up with a tool to do this > kind of searching > >> and indexing, and the query of the index could be > run on your local > >> node? The indexes themselves could be distributed > as content within > >> freenet (so you're not compromising anonymity by > using them) and since > >> the indexes are inserted under a certain key, > they'd be no more or > >> less vulnerable to poisoning than trusting one of > the existing index > >> sites. Instead of going to tfe's page each day, > you'd grab your search > >> index from tfe and perform your keyword search on > your local node. > >> Am I missing something obvious here? > > > > That is the way to go, when things get big. More > convenient than the > > sites perhaps, but not more secure. Although you > could combine several > > anonymously inserted indexes using one client. Oh, > and it's likely to be > > pretty slow unless the indexes are so small that > you can fetch the whole > > thing every day. > > Combining multiple indexes into one search client > makes a lot of > sense. It would lower susceptibility to a particular > indexers bias > (either intentional or algorithmic) and it would > help with the > "indexes get big" problem if you could fetch large > indexes once > (in a while) and incremental indexes most of the > time. > > Now all we need is an small, easily mergeable, > quickly searchable, > standardized index format... and the tools to go > with it. :) > > -michael > _______________________________________________ > devl mailing list > [EMAIL PROTECTED] > http://hawk.freenetproject.org:8080/cgi-bin/mailman/listinfo/devl __________________________________________________________________ Gesendet von Yahoo! Mail - http://mail.yahoo.de Logos und Klingelt�ne f�rs Handy bei http://sms.yahoo.de _______________________________________________ devl mailing list [EMAIL PROTECTED] http://hawk.freenetproject.org:8080/cgi-bin/mailman/listinfo/devl
