Why are search engines so different

The Jux2 Meta Search engine which provides results based on the results of google,yahoo and MSN provides a rather surprising statistic showing how much these major search engines differ in the top queries results.

Is it because the indexing method is different for these search engines? Or just a result of the type of use by the users of the searchers! I mean, do some results move up due to that being clicked upon by many users.

The rate of re-indexing will also be different, which can lead to the differences in the results. All this aside, the algorithms used by these engines should be different leading to these differences.

Coop search is now on yubnub

Using the Yubnub command line, the required search engine can be found. To find a coop search just type"fcop query"

So to find music CSE's just type fcop music and it will show u the list of all music CSE's.Try it by clicking on fcop music.If you have not tried Yubnub, give it a try, its useful and so much like the Unix shell, u don't need to learn anything new.

Yubnub has got a huge collection of commands and u can add ur own too.

Veronica-a search engine for gopher protocol

The Gopher protocol is now used very rarely if at all. But Veronica was one of the first indexing engines. The latest version known as(Dont click on this link with internet explorer-use firefox or other browsers that support the Gopher protocol) Veronica-2 is still active. It is of interest as it does not index the contents of the files.

If you click on Search Veronica-2, a dialog box will appear where you can provide your query. The list of gopher servers that have the word(s) will be listed.Try using "SRS" as the query. It refers to the the servers with protein and nucleic acid sequences.

The latest addition to the search engines is VISHNU(Search VISHNU (experimental)), it arranges the results based on the folders in which the files are stored. It also shows the number of files in that folder. To begin with the indexing engine should have these basic features. Better features can be added later.

Asian Cinema Search engine

Search engine Topics:Asian Cinema

Contribution: [Volunteer to contribute]

Results hosted at:Google

What's Interesting:Its claim to fame "Asian Cinema Search includes sites with coverage of East Asian cinema. Whenever possible, sub-sections of individual sites are included with separate tags to make search refinements as focused as humanly possible.".

You get a large number of result refinements. Refine results for doom:
Reviews		News		Trailers		Forums
Blogs		Wikis		Databases		Links
Distributors		Official Sites		Interviews		Articles

You can compare the results from the Japanese Cinema CSE with google.

Try it:Asian cinema search

Reviving the meta tags

Meta tags have been used for describing a website and its content. But the use of meta tags has declined so much so that they are considered dead. Very few sites seem to have retained the meta tags on their sites.

The new buzz word is bookmarking. So why not use meta tag data for bookmarking?

The main argument against meta tag content was that it included more keywords than the data contained in the site. The idea of bookmarking is that the user makes sure that the site actually contains that data. The user has to type in all the tags before bookmarking. Here is a greasymonkey script to find meta keywords and description and provide a simpy this link on top of the page.

Simpy this meta data

The user can choose which keywords are appropriate and which are not. And the user is spared the burden of having to type in the tags and description.

Google results with a time line

Google has introduced a new experimental method to list results using a time line. The results such as this one on "nanotechnology"will show the results based on the word and the date it is related to.

The idea works fine with the examples they have given, but for some of the newer words created post google, they better use the information as to when it was cached. If the cache is not backed up, its hight time they started recording the dates on which new words were found by the google bot. This has many advantages to the casual linguist and also provide a more reliable time line for the future.

I wonder whether the site map can have data which can be used to build the time line. The map view is more of a browser version of google earth.

Mahalo-human powered search

Recently a search engine "Mahalo" was launched in alpha version. Its claim to fame "Mahalo is the world's first human-powered search engine"

It makes use of "Guides" who are human to sort through content and remove spam. I came across a guide by name "Nicole Gustas" who has created as many as 227 results!! Converting that into google coop's will mean nearly 227 search engines.

The results are marked as "Warnings , Guide's Choice , and What is? icons". Although google coop does not allow for showing warnings sites can be excluded. "What is" seems to be an equivalent of the answer.com results shown in Google, may be slightly better.

The rules for selection are rather stringent..They don't want any sites that have too many advertisements, sites which are younger than one year and so on.You can find details at their FAQ.

I compared the result for the word "Advertorial" from Mahalo with that from google . Although the first result is from wikipedia in both, which reminded me of Search engine nightmare. The search refinements in Mahalo are rather unique, the refinements provided by google dont recognize the context as well as mahalo as they are "not human crafted". May be a better algorithm would do the job for google.

But how do they plan to generate money from this endure? by selling the search results to search engines or by ads or the other means?

I almost forgot you can create an account and add sites, suggest sites to be included for a particular result.A lot of similarity between mahalo and google coop.