Slashdot Mirror
New Search Engine Cuil Takes Aim At Google
theodp writes "CNET reports that Cuil (pronounced 'Cool'), a startup founded by the husband-and-wife team of Xift creator Tom Costello and former Google search architect Anna Patterson, is launching a new search engine today that claims to index three times as many Web pages as Google." Running a few searches left me underwhelmed with the content of the results (hitting the next-page button on a search with a listed 62,200,000 results — for "seattle" — got me the unexpected error message "We didn't find any results for 'seattle.'"), but pleased with the actual layout of the results when it worked, so I hope the kinks are worked out. Update 7/28 18:30 GMT by SM: corrected Tom Costello's accreditation, he wasn't a professor at Stanford as the linked story suggests, just did some research there as a grad student. Thanks to the Stanford CS department for pointing this out.
A few observations:
* "Cuil" is a really dumb name. "Google" is a dumb name too, but at least its pronounciation is obvious to anyone reading the name. Can't say the same for "Cuil".
* It's unlikely that this new search engine even approaches Google in its comprehensiveness, or ever will
* Cuil has some weird bugs. I searched for my name, found a link to a Gallery page I have about my son's birth earlier this year, and they have a little thumbnail icon next to the search result for that. But it's a random map of the United States completely unrelated to the page it links to. Bizarre.
* Cuil's results come back more slowly than Google's (but this is from New Zealand, maybe it's faster from the USA), and their page re-renders in odd ways (at least on my oldish Firefox install) as results come in.
* Cuil seems to give the most favor to any page that has the word "is" after the search term. Invariably, the first result for almost any single word search will be whatever page starts out with "[Search Term] is ...".
* Google is really bad for Silicon Valley. So many good software developers in SV got sucked in by Google. Too much of the top talent in the area is now working for Google, doing almost completely useless stuff, and it's not healthy for the industry. Is there any software company in the bay area that hasn't had at least a couple of engineers sucked away by Google? Are algorithms for pushing targeted ads and useless web applications that never get out of beta really worth depleting the industry of so many of its best? I predict that when Google comes crashing down (and it will - anyone who has seen the ridiculous excess of the Google campus cannot help but realize this), the net result will have been to set back innovation in the software industry a great deal, by tying up so many people who would otherwise have done something useful.
* For the above reason, I wish Cuil all the success possible, because I'd love to see some actual competition in the search engine world.
Anyway that's how I see it.
And I typically got relevant results with little spam, but that may depend on what you are searching for.
I initially got abysmal results (no result found for just about anything I searched for, like the very technical "implicit volumes" or "queyras")
Then I deactivated "safe search" and finally obtained some results. However I suspect my original good impression I had of having found "relevant" (authoritative?) results in the first place were due to the safe search being on.
. . . . . . .
may u!sh 2 sm!le at dz!z bad nn.!m!tat!ion
.....and then shut down -- if the servers overload that quickly, they're going to fade into obscurity pretty damn fast.
But I saw enough in those 5 minutes to realize it has major problems:
-- Returned fewer results than Google on 2/3 of my searches
-- Compound "words" (such as, say, "georgebush" as opposed to "George Bush" as might be found in file names, tags, captions, etc.) produce NO results
-- Eliminates common words, connectors, and even pronouns from exact phrase searches, which defeats the whole point
-- Seems to have no provision for ordering by date or viewing most recent additions
-- Also does not seem to allow more than 10 results per page, which severely slows things down
-- Their "safesearch" (which I wouldn't use, but I wanted to try it for comparison) seems to eliminate even some innocuous terms
-- Some of the images that accompany the entries have nothing to do with the actual webpage listed
-- The layout sucks with their "paragraphs all over the page" format -- Give me a LIST, dammit, that I can quickly scroll through and scan
Overall, just as useless to me as most search engines have been. Google has its own faults, and you may rightly criticize them for their ethics, privacy policies, or business practices, but it is still a far better tool, and no one is seriously challenging their dominance anytime soon.
UPDATE: It just came back up. Why am I seeing many of the SAME results on page 2 and 3 and 4 as I saw on page 1? Is it just repeating entries to inflate the numbers for results? This thing is not just "not ready for prime time" -- it's not even ready for "obscure middle-of-the-night cable slot."
"Every great cause begins as a movement, becomes a business, and eventually degenerates into a racket." -- Eric Hoffer
They also mention
If you have modified your robots.txt file for Twiceler, it may take several days for us to re-read the file. If you need something blocked right away, please let us know.
Unfortunately it looks like they don't search in non-english domains. All my Polish-related queries gave 0 results. Pretty much useless to me now.
Their privacy policy gets a thumbs up:
...when you search with Cuil, we do not collect any personally identifiable information, period. We have no idea who sends queries: not by name, not by IP address, and not by cookies (more on this later). Your search history is your business, not ours.
We do not keep logs of our users' search activity.
We do not record the information in your cookies on our servers; your browser sends your preferences to us with each search request. This way, we do not store any personal information about you on our servers.
Comment removed based on user account deletion
Must have been fixed amazingly quick. I have no problem using it now... =)
/.Mattsson - My native language is not English, so please don't whine over linguistic errors. (That's lame anyway...)
Yeah, thanks for maintaining old biases. I'm sure the Aussies get as much grief, but really let's keep that kind of borderline racist attitude off the table and stick to the intellectual and geeky stuff
Interesting as a fluent Irish speaker I am not familiar with Cuil or CÃil meaning knowledge or wisdom or anything similar. It does however mean curls, a goal (as in sports) or occasionally someone's behind - as it does in many languages. The thing is that leaving an 'i' in it means it would be genitive - not a standalone word but part of a reference or possesive case e.g. cÃl, mo chuid chÃil.
Back to the Irish/Aussie thing, a lot of the words you love and know as Australian are in fact rooted in Irish. Let's not forget that Australia was a prisoner colony and Irish being one of the biggest nuisances to the British Empire at the time, we tended to make up a sizeable portion of the population.
I never get used to these constant resurrections
This
You just got troll'd!
http://www.cuil.com/info/
Cuil: "three times as many as Google" = 121,617,892,992
http://googleblog.blogspot.com/2008/07/we-knew-web-was-big.html
Google: "1 trillion (as in 1,000,000,000,000) unique URLs on the web at once!"
Seems like Cuil has 12% the links of Google, not 300%... seems like more than a small error.
Have you tried Clusty? I gave it a go as my default search engine a while ago. It consistently ranks the search results much better than Google but, again, is hampered by the fact that it's database is much smaller. My desired result might be on the fifth page with Google, but it won't be in Clusty at all.
I am TheRaven on Soylent News
"zeta potential analysis of gold colloids" returns no results whereas google returns 44,300.
Yeah, that's kind of disturbing. It shows pictures of all kinds of people next to hits with my name highlighted. It also finds images that I've made, but associates them with random pages other than the ones they actually appear on. Looks like they need to refine that, because that may be turning people off. But it's not a bad idea in principle.
EagerEyes.org: Visualization and Visual Communication
Funny... I saw a pic not related to him as well when I clicked on your search. Unlike my earlier test searches, clicking on the 'next page' link actually worked (somebody already mentioned the problem I was having: no search results found despite several thousand being available according to the first results page)... which led to another pic-problem: the wrong party logo associated with him! LOL
Since I comment a lot on various blogs, I've found it's easy to keep track of what I said and where by just googling the name I use. Google gives me very relevant, easy to understand results with very few false-positives (unknown misses, but it seems good in that respect). Cuil, on the other hand, was a bit off. The results look like they're all me in that the sites are places I've commented at, but the text blurbs for the links are collections of pieces of comments, some mine, others not. There are a lot of repeats. The pictures that come up have nothing to do with me and the ones I've checked has nothing to do with the item linked to.
Speaking of links... I'm sure it's a matter of getting used to how Cuil does things, but the linking is very much NOT obvious. All these links that are supposedly to my individual comments are each for a given story, right? Nope. The links at the top and bottom of each search result are to that website's log of my comments. So I have a bunch of links to my Lifehacker member's comments. Another to Gizmodo's index of my comments. (Yes, they're the same system, so it's the same list with a different background, but that isn't any search engine's fault.) I even get this for sites I've only left one comment at. Yet, each search result claims (via its headline) to be linking the story or comment directly. The only time I've found that to be true is on the websites that don't provide an index of members' comments (like Popular Science).
I want the competition and I hope it improves. The layout is pretty nice (well, would be if the info was accurate). So far, color me unimpressed.
I hope this comment is well received... I could have moderated instead!
Persecutors will be violated!
Wow! Just had a go on it! That is one slow search engine. I know it will speed up with less people trying it out, but it cant be getting any where near the traffic of Google, and it is incredibly slow.
Yeah, that's funny coming from somebody on slashdot.
Whatever.
http://slashdot.org/faq/slashmeta.shtml#sm150
"Slashdot" is a sort of obnoxious parody of a URL. When I originally registered the domain, I wanted to make the URL silly, and unpronounceable.
2784 is an integer multiple of 3, so it's more like 1 in 3, but then it's unwise to generalize from a data set where n=1.
Bzzzz wrong. For the result of a division by 2784 to be an integer, the numerator has obviously to be a multiple of 2784. Therefore I'm right in saying 1 in 2784. I accept your apology.
You just got troll'd!
in what way is it difficult to pronounce nintendo? It's said exactly the way it is written.
Not only is their name hard to remember, its pronounced the same as a totally different word.
Fail
Fail.
DRM-free indie games for the PC and Mac: Positech Games
http://www.cuil.com/info/privacy
What changed under Obama? Nothing Good
I've done a number of simple searches on both Cuil and Google to see how the results compare. Despite Cuil's claim to have more indexed pages, Google is consistently giving me a larger number of results, and they tend to be more useful, as well.
Someone needs to tell these cuil guys that it's quality, not quantity, and they're currently failing at both.
Those who advocate genocide deserve every protection afforded by law, and none afforded by common human decency.
The algorithm isn't just bad, it's horrible. I guess it's halfway decent for popular news items and such, but with the first technical term I tried ("Gödel's incompleteness theorem", in this case), well, the results are not looking so great. Compare:
Google vs. Cuil.
I mean, you don't even have to know what the theorem is to see how much better Google's results are for someone who'd be searching for basic information on it. I mean, for Chrissake, Cuil returned a bunch of garbage from what looks like a Christian theology forum on its first page. NOT RELEVANT
If you want to beat Google, you need something new. Natural language searches, search engines that act as agents continuously looking for things for you, whatever.
Like Google Alerts?