Will Napster-style peer-to-peer searching mean an end to search engines? Despite the continuing hype, I doubt this will be a replacement for web-wide searching. There are strong advantages to a centralized system, not the least in terms of dealing with spam. I do think P2P has promise for intranet solutions, especially in situations where you have trusted data that you want to find on a few hundred or thousands of machines. But P2P for full-text search against a billion or more URLs? Unlikely. Below are two new articles, pro and con on P2P search, plus my last article on this topic.
Search project prepares to challenge Google
News.com, Feb. 26, 2001
Pandango is an interesting idea. By examining what we visit, it would then select the more popular documents in response to a search. This is akin to the bookmark-based search engines that were all the rage this time last year but which have failed to take off. It's also somewhat similar to the Alexa "Related Pages" technology already built into Internet Explorer and Netscape Navigator but which I've yet to encounter many people using. Aside from the big disadvantage of requiring users to download software, there could also be privacy concerns. Long story detailing the project. Despite the company's suggestion they are the "third generation" of web searching, when Pandango launches, it will really be a first generation service. The technology may be a move in a new direction, but until it has been tried and tested against a web-wide audience, it can't really grab that claim to be the next rung on the evolutionary ladder of web search.
P2P Goes in Search of 'Doogle'
Wired, Feb. 16, 2001
P2P is unlikely to replace centralized web search engines, is the conclusion from this major P2P conference.
More Than Just Music Search
The Search Engine Report, June 2, 2000
Napster has obviously changed since this was written, but the contrast and comparison of centralized web search to the potential of P2P search remains relevant.