John Battelle's Searchblog

Hmm... top 100 by what measure? I see many notable vertical misses, including (my bias) Healthline, http://www.healthline.com (heck, I was even on a Search SIG you moderated John ).

Nontheless, a nice who's who!

This would be a perfect topic to use the SNAP previews with

http://www.snap.com/about/spa1A.php

See if perhaps it could be intergrated with this topic; the thumbnails of those various search engines would be displayed as one hovers over each link

I dont think its a toplist based on traffic, if so its highly unaccurate.

What about http://www.compete.com and alexa? I'm not digging this list.

So many search engines, so little measurement of text thinking. Coning complements search. See www.oracep.com

Is digg really a search engine? Maybe we need a definition of a search engine.

Like: A Search Engine is any site that has some way to request a specific piece of data.

Great List.
I see several social bookmarks sites?
Amazon? a search engine?
This a very interesting list is will review further.

"This would be a perfect topic to use the SNAP previews

See if perhaps it could be intergrated with this topic; the thumbnails of those various search engines would be displayed as one hovers over each link"
Good idea

I wonder if, during the first round of search wars (AltaVista, HotBot, etc.), anyone devised a relatively unbiased way of measuring search accuracy. It would be interesting to see a metric for how often an engine returned what the user was looking for on the first page.

Greetings everyone.

This is my Top 100 list, and the criteria I used to start was "that looks interesting." - in other words it was subjective and not scientific. But having said that:

I appreciate any and all feedback to the topic(s): The Top 100 Search Engines for 2006. The Top 10 (out of those 100).
The #1 Search Engine for 2006 (out of those 10). (I picked ChaCha) and finally the Top 10 Search Engines to watch in (early) 2007.

I will read and review every comment and use them to improve the list.

Thanks to everyone.

Charles Knight

I can't believe you omitted Dogpile and Info.com.

Dr. Pete writes: I wonder if, during the first round of search wars (AltaVista, HotBot, etc.), anyone devised a relatively unbiased way of measuring search accuracy.

Pete, such relatively unbiased ways of measuring search accuracy have been around since the 1960s (Cyril Cleverdon and the Cranfield experiments). Those measures, utilizing concepts such as relevance, precision, and recall, are still used today with little change.

The interesting question, IMHO, is not whether these measures exist. The interesting question is why the major search engines won't submit themselves to a public, TREC-like evaluation of their algorithms. For example, let's take 5000 queries, sampled randomly across popularity (high, medium, and low traffic queries) and across topic (computers, fashion, travel, etc.) Pool the results from the participating engines (GYMA and whoever else wants to participate) and spend a couple of trinkets from that $150 billion market cap to pay an independent third party to assess the relevance of, say, 50 documents for each query.

Then it is simply a matter of using recall and precision to see which engine, on average, produces the best results!

Oh, now THAT would be an innovation! Public evaluation!

Of course, it makes sense why some would do this, and others wouldn't. Public perception already rewards some engines as better than others. An engine that is already a popular favorite really does not benefit from such public evalution. If it wins the evaluation, it garners no more favor than it already had. If it loses, or even just ties for first, then it knocks the engine down a peg, in public perception. So there is absolutely no incentive to participate.

And yet, from a user's standpoint, public, 3rd-party evaluation would be an innovative godsend. The public might even learn that certain engines are better than others, at particular topics! One engine might do really well on computer queries, another on fashion queries. One engine might be really good at handling popular queries, and another much better at long-tail queries.

I would love to know which engine works best in which scenario. So, c'mon GYMA.. how about some consumer-oriented innovation in that sphere! I could care less about all your chat clients. How about some public evaluation innovation, instead?

If we are going to call Amazon a search engine why not eBay or Wikipedia or Myspace for that matter?

Interesting list but I'd love a bit more information on criteria and qualification.

I like http://pizza.de -- but I guess perhaps the difficulty of getting a hot, steaming delivery to the States might arguably disqualify it from Charles's list. You can't always get what you want....

;D nmw

JG's thoughts on public evaluation are sound, but how does the public evaluate a result? It looks for meaning and relevance. Coning algorithns rate and rank the meaning in searched results. This gives relevance genuine thinking credibility. For more, see www.oracep.com and contact us for sample research.

I note Quaero isn't on there. Heh.

Just completed reading "The Search" John.. Great read and a Great Book. Thanks for all the effor you have put into it.
- just wanted to point out here that you have lfet Yahoo Answers here, in the list of search engines, i can see Live QnA but not Yahoo Answers

@JG - I'm glad to know that at least the research is out there. I'm a cognitive psychologist by training, and I'm amazed at how often no one bothers to even think about testing something that seems so readily testable. Even if the results of applying such testing to the major engines weren't completely fair or accurate, it seems like they would kick-start a fascinating public conversation about how an ideal search engine should function.

Does anyone have the link or URL for Quaero? Is it live?

Here are JGs thoughts Coned; public evaluation begins!

Coning Index 90% -low level background, high level analysis and judgement.

66%] Pete, such relatively unbiased ways of measuring search accuracy have been around since the 1960s (Cyril Cleverdon and the Cranfield experiments). Those measures, utilizing concepts such as relevance, precision, and recall, are still used today with little change.

93%] The interesting question, IMHO, is not whether these measures exist. The interesting question is why the major search engines won't submit themselves to a public, TREC-like evaluation of their algorithms. For example, let's take 5000 queries, sampled randomly across popularity (high, medium, and low traffic queries) and across topic (computers, fashion, travel, etc.) Pool the results from the participating engines (GYMA and whoever else wants to participate) and spend a couple of trinkets from that $150 billion market cap to pay an independent third party to assess the relevance of, say, 50 documents for each query.

100%] Then it is simply a matter of using recall and precision to see which engine, on average, produces the best results!

100%] Oh, now THAT would be an innovation! Public evaluation!

92%] Of course, it makes sense why some would do this, and others wouldn't. Public perception already rewards some engines as better than others. An engine that is already a popular favorite really does not benefit from such public evalution. If it wins the evaluation, it garners no more favor than it already had. If it loses, or even just ties for first, then it knocks the engine down a peg, in public perception. So there is absolutely no incentive to participate.

100%] And yet, from a user's standpoint, public, 3rd-party evaluation would be an innovative godsend. The public might even learn that certain engines are better than others, at particular topics! One engine might do really well on computer queries, another on fashion queries. One engine might be really good at handling popular queries, and another much better at long-tail queries.

75%] I would love to know which engine works best in which scenario. So, c'mon GYMA.. how about some consumer-oriented innovation in that sphere! I could care less about all your chat clients. How about some public evaluation innovation, instead?

Want to know more? www.oracep.com

Honestly how useful is this list? Maybe 5 of these guys even matter...

Am I blind, or are you telling me that live.com isn't part of the list? you have the QnA, but not the main site?

How about travel search engines such as www.mobissimo.com?

This site offers totally free search engine submission tools. PromoteLikeCrazy.Com

Quite impressive to see Technorati being more successful than amazon at search. I think the clean interface and the clear value to the user are the main reasons for technorati's success.-

Thank you for sharing this story with me !

FREE MONEY SEARCH E-BOOK
The best kind of searching is for unclaimed funds!

FREE E-BOOK

Just want to tell you that yoople is linked to a wrong address. Yoople! - Collaborative Web Search is at www.yoople.net nope dot com!

http://www.simplemw.orgfree.com

I've found ClipBlast.com the best Video Search to get to video beyond user generated sources.

Great list!