A new type of search algorithm is described at NSU. In a way, it is the next generation over Google. It works off the principle that most web pages link to pages that concern the same topic, forming communities of pages. Thus, for academics, this would be great as the engine could find the community of pages related to a certain subject. The article also points out this would be good as an actually useful content filter, compared to today's text-based ones
Sounds like PR taken to a higher level. But not looking at the actual content of the pages...hmmm.
Sure a good approach in academic / non commercial area, where PR also worked good.
This one looks fine for a research engine, but as WebGuerrilla said elsewhere, some algos would likely find life very tough outside academia.
>find life very tough outside academia
Agreed. I think of them as "Scooby SE's" as in...
"l would of gotten away with it, if it wasn't for those pesky kids and that dog"
There are plenty of us "pesky kids" about, they are building the algo the wrong way round. imho
*lol* NFFC that has to be the best analogy that I've heard :)
When I first read it I said,
Scooby fits it better.
The term `cargo cult' is a reference to aboriginal religions that grew up in the South Pacific after World War II. The practices of these cults center on building elaborate mockups of airplanes and military style landing strips in the hope of bringing the return of the god-like airplanes that brought such marvelous cargo during the war. Hackish usage probably derives from Richard Feynman's characterization of certain practices as "cargo cult science" in his book "Surely You're Joking, Mr. Feynman!"
Jargon File 4.0.3
Anyone want to compile a new SEO glossary? ;)
(edited by: digitalghost at 10:32 pm (utc) on Mar. 7, 2002)
I guess the achilles heel is the "community behaviour" aspect.
Design an algo based on milquetoast, and the thugs take over.
Oh well, I guess that is why we have law enforcement ( zero PR).
Maybe it would be better to just ship the thugs off to their own island.
>why we have law enforcement
I think it would take a whole lot of policing for that one, probably enough to break the budget.
>When I frist read it I said
When I first read it I drooled.
The problem seems to be that they designed it for responsible webmasters rather than the lowest common denominator, Spammasters.
The communities they speak of do exist, but the interests within the communities are diverse. The diverity is reflected in the links as well as the community interest.
A simple web development site might have links to XML, .asp, DHTML, Curl, Java, CFM, PHP, Perl, etc. Those all based on theme content. Affiliate links might point to cigars and sofware to pay the bills.
A site entirely about XML links to the web/dev site and reviews XML books. A site that provides links to book reviews links to the XML site. A site selling crafts links to the book site. That's a community alright...
Content is more than links, and using links as the sole determination for relevance won't cut it.
I have to assume that that article was just poorly written. How on earth can you have a search engine that ignores the text on the page? Text has to be figured in somehow, unless it is only going to return category type matches. The way it reads, it sounds like it would just be a really good directory.
The other point the article misses is the fact that there is already an engine that looks at links and identifies communities. It's called Teoma, and it is doing a pretty darn good job.
If you do a comaparative search for the term search engine optimization on Google [google.com] and Teoma [teoma.com] you can see a clear example of the differences.
On Google, the number one listing is Paul Boutin's article on SEO. On Teoma, it doesn't show up because Teoma understands that wired.com isn't a prominent site within the SEO community. What is also interesting is the fact that Teoma returns both WebmasterWorld and I-Search in the first page, despite the fact that neither page really contains the search term.
Maybe I'm missing something, but it seems to me that Teoma's approach is doing a pretty good job of addressing the issues the guys at Princeton feel are being overlooked.
How many different ways can the .edu's interp Authoritative sources in a hyperlinked environment [cs.cornell.edu]? That must be the only text book they are using in cs data retrevial 101 these days. call me when Cornell puts out a search engine.