On Friday 15 August 2003 14:03, Newsbite wrote: > I don't know why some people claim it's not possible to create a true > searchengine; I've already got information one is working on something with > frost. And tho I doubt it's going to be what amounts to a 'normal' > searchengine, I do remember reading a paper here about a Scalable Anonymous > Freenet Searchengine. It was in pdf format, I believe. Maybe Ian can shed > some light on this or paste the link/paper again.
I read that paper. Unfortunately, IIRC, the proposed architecture is inherently contradictory to the primary goals of the Freenet design. > so at least in theory, it's quite possible to create a searchengine for > non-centralised networks. Almost certainly. But it is not an easy problem to solve. > Besides, though some arguments are valid, I think some are too pessimistic. > If one takes the indexes as a group of keywords and/or the short > description + the link (and thus not the duplication of the content), then > every single entity in the database would stay under 5kb. So, thousands > could be placed in the database, before the js got too heavy for practical > use. 1) An index that is not based on all of the content of a page is nowhere nearly as useful as having the index of all ther terms on a page. 2) You seem to vastly overestimate the capabilityes of JavaScript. Remember that this language was originaly designed to deal with things like roll-over graphics. Now you are talking about creating an entire client-end application in it. While it may be possible, there are issues that may be difficult to overcome... Gordan _______________________________________________ devl mailing list [EMAIL PROTECTED] http://hawk.freenetproject.org:8080/cgi-bin/mailman/listinfo/devl
