A new large search engine called Cuil (pronounced “cool”) was launched this week.
By Pandia Guest Writer Lars Våge, Internetbrus
According to Cuil the search engine is three times as big as Google, Yahoo and the others, with more than 120 billion web pages in its index.
The last time I say figures this high in relation to other search engines was for Recall, the short lived full text search of the Internet Achive in 2003. (That being said, Google now claims to have found 1 trillion pages on the web).
Anna Patterson from Stanford University was responsible for the Recall Project, and she is actually one of the founders of Cuil.
In this team we also find a certain Mr Louis Monier. Yepp, that was the man that created AltaVista. Both Anna, Louis and some other members in the Cuil team have until recently been working for Google.
Does size matter?
So to what extent does the size of the database mean anything for search engines?
Given that most of us seldom go beyond the first page of results and its 10 links, it does not mean that much. Only very specialized search queries (or very badly constructed search phrases) deliver such a small number of results that you can really compare results between search engines. So far I have failed to see examples of Cuil delivering more hits than the others.
As regards the relevance of Cuil’s search results, it is decent, but not particularly impressive. There is definitely more work to do for Patterson, Monier and the others in this area.
In the press release they talk about “content-based results, not just popular ones, providing different and more insightful answers”. If that implies that they mean to skip link analysis altogether, they should probably think again.
Anyway it is always exciting to test a new big search engine.
User interface
The user interface is a bit more engaging than the others (Ask.com excluded). Instead of a linear list containing 10 hits and the odd addition of a YouTube video — like in Google — we get three columns of hits.
Each result includes a text extract that is longer than the ones we find among the other major players, and there are pictures to illustrate the search results. This means that there are fewer results visible at the same time, but the result page look more attractive.
Strangely enough some of the images are not from the relevant web site. For a search for our site internetbrus.com, one of the results were illustrated with a rather attractive butterfly pattern!
The background color of Cuil is pitch black (meaning that the layout is energy friendly) and the other colors are ultramarine and grey. The area with the results have a white background.
Interesting features
There are a couple of features here worth mentioning. As in Yahoo and Ask you get suggestions for queries as you are writing in the search form, which may help you save time.
Cuil will also suggest alternative and more focused queries in the search engine result list. These alternatives are presented as tabs. When searching for Sundsvall, I got tabs for the major sport teams Gif Sundsvall, Sundsvall Dragons and Sundsvall Hockey. That’s not bad considering most of those pages are in Swedish!
Explore by Category
Then there is the Explore by Category feature. At first I believed these were automatically generated clusters, but instead Cuil seem to be using some kind of named entity recognition.
What makes it different is that you not only get recognized names of persons, places or events found in the text of the relevant web pages. They are actually divided into meaningful categories like bebop-pianists, Indian cricket players, Swedish football league clubs, composers writing for pipe organ and so on. I have not seen anything like this before.
If you click on a link in the Explore by Category window it helps you focus your search, as the category is added to your search query. This is similar to what happens when you choose to focus your search using Cuil’s alternative search phrases by clicking on the tabs.
Privacy
There is unfortunately no documentation of the search syntax available as of yet, but that will probably be added at a later stage.
Cuil does not keep any information in search logs that can be used to identify the searcher. In other words: The people behind Cuil seem to be aware of the stronger demands for search engine privacy that we nowadays find for instance in the EU.
We’ll have to wait and see what happens to Cuil and if they are able to improve the weaknesses mentioned above. Regardless: It is inspiring to see that there are people out there who wants to give us new search alternatives.
This article was originally published in Internetbrus, a Swedish blog on search engines and Internet searching that has been online since early 2001. It is written for both searchers and educators.
Internetbrus is owned and edited by Lars Våge and Lars Iselid. Lars Våge works as a librarian at Mitthögskolan and a programmer for JL Informationsteknik. Lars Iselid is a librarian at the Umeå University Library, freelance journalist for the computer magazine Datormagazin. He can be found blogging under the pseudonym Cyrille at Iaslash.org.
Lars and Lars are co-authors of a book on Internet research: Informationssökning på Internet.
Original post by Per and Susanne Koch