cuil - A new search engine

D-A-L · #11 (**permalink**) 29-07-2008, 09:47 PM

What a waste of publicity. They really could have done something special and taken a ride on the media wave to just possibly have a bite at Google. How can a bunch of people with that track record get something so wrong? How on earth did it make it past even early testing? Surely you benchmark your search results against the best in the business, Google, and as soon as you do that its a non starter. Weird. Almost like an april fools joke.

Digerati · #12 (**permalink**) 29-07-2008, 11:05 PM

It wasn't just publicity - it was boasting with great fanfare! I wonder if they had a PR firm set up the announcement and press release.

I think it will be interesting to see how they attempt to recover. Taking care of bandwidth is easy, if you have the cash. And it seems the more they see what the public is searching for, the more they could fine tune the process.

If the 9 results on the results page consistently gave me the answers I was looking for, I might be persuaded to use it.

My problem right now with Google is I am not sure if the top result is what I want to see, or what Google wants me to see.

rokytnji · #13 (**permalink**) 30-07-2008, 04:19 AM

Well tried out Cuil. They had a pretty picture of a girl looking at her laptop screen in the grass next to the DAL link. I checked there preference and privacy sections, liked the safe browsing options. Also liked that they aren;t going to share my ip address. That alone convinced me to bookmark their browser. I don't know yet if I'll make them my home page or default browser yet. Too soon to tell but I like their mission statement. Time will tell I guess. Oh and it worked faster than google on my Amrel laptop.

jephree · #14 (**permalink**) 30-07-2008, 05:26 AM

I haven't found the interest or time to try it but have been reading about it:

Wiki says:

Quote:

Google began in January 1996, as a research project by Larry Page, who was soon joined by Sergey Brin, two Ph.D. students at Stanford University in California.[6] They hypothesized that a search engine that analyzed the relationships between websites would produce better ranking of results than existing techniques, which ranked results according to the number of times the search term appeared on a page

And Cuil says:

Quote:

Popularity is useful, but not always important

Popularity is useful, but has dominated search results so heavily that it gets harder and harder to find the page you want, especially if your search is a complex one. Cuil respects popular pages and recognizes that for many simple searches, popularity is an easy answer to your question. But for a deeper search, establishing relevancy is more than a numbers game. Cuil prefers to find all the pages with your keyword or phrase and then analyze the rest of the content on those pages. During this analysis we discover that your keywords have different meanings in different contexts. Once we’ve established the context of the pages, we’re in a much better position to help you in your search.

And BrickSalad via Reddit says:

Quote:

So, apparently what cuil is trying to do is make a hybrid of google's popularity ranking techniques and an advanced variation of pre-google techniques. It's not a bad idea, but apparently the implementation of the idea was a complete failure (just do some searches to discover this)

More Reddit posts and comments follow here: NSFW / Adult Material / etc.

Quantum Porn: Cuil Reinvents Web Search (NSFW) : reddit.com

Ex-Googlers reinvent web search | The Register

jephree · #15 (**permalink**) 30-07-2008, 06:21 AM

It does make for a lot of interesting reading:

I don't understand what Cuil is doing : reddit.com

Quote:

That's the problem, and the index seems to be building. For stuff that is indexed, the results seem refreshing.

Since the poster says he doesn't understand what they are doing, I'll try to explain the idea of indexing. This is in general terms, and someone may have a better explanation than me.

First of all, searching or sorting data is about the slowest thing a computer can do. There's all sorts of theory on how to build the fastest search algo. What Google does, is not build a search algo, but an index. So, the searching isn't done in real time, but some time before you search for a term.

For every term someone would search for, the results have to be predetermined. If they, or Google, tried to determine the results at runtime, it could take hours to return one query. So, every page it spiders, it archives. It then determines the relevant terms. Then, when an index is being built for that term, it will generate a list of maybe the top 10,000 articles related to that term. Then, when you search for it, you get that list in order, which is very fast and easy to return.

So, you search on this engine, and the results for every word in the dictionary do not yet exist.

I'm not sure how clear that is, but I've done similar things before, and it takes lots of space to hold the index. It also takes a lot of time. I understand, that in the early days, it would take Google a month to rebuild their entire index. Currently, Google is updating their index on a continuing basis.

So, they have articles related to COBOL, but no index for it. That, or it's a joke, because the use of COBOL cripples the mind; its teaching should, therefore, be regarded as a criminal offense.

D-A-L · #16 (**permalink**) 30-07-2008, 08:24 AM

They should never have released it so early. They are getting a lot of negative PR right now which I guess could be a positive thing in terms of the fact we are all talking about them but if nobody uses the service then in a few months the name could be forgotten.

Bottom line is people use a search engine to find the most relevant search results for their query. After that they worry about Privacy. I was reading yesterday someone saying that instead of creating a search engine, if they are so worried about privacy they should have developed a tool which protects your privacy when searching.....

k_9 · #17 (**permalink**) 21-08-2008, 08:47 AM

Quote:

If nobody uses the service then in a few months the name could be forgotten.

I think i might be using it. Just got round to having a look at it and i'm quite impressed by first. I like the way they layout the searches and give a lot more information than google before you actually have to browse the site. I also like the box wich gives an individual acknowledgement of the testemonials, and that there are people online to help you every single day. Compared to google, it's a lot more interesting, but the name at first when i saw you talking about it put me off. I haven't yet found any problems which i'd class as "problems" and the privacy is also a plus. If they could get more people to know the name, then maybe, just maybe it could become a success.

Digerati · #18 (**permalink**) 21-08-2008, 03:27 PM

I keep going back to give it check to see if I will change my mind (again).

I still have mixed feelings about it. When it gets a solid hit on your search string, I like it, but if it has to rely on "fuzzy" logic, it misses badly and behaves oddly. For example, I am restoring my first audiophile quality speakers, a set of Acoustic Research AR3a Speaker Systems. They cost me a month's pay (each!) but I was living in the barracks on a rock in the middle of the Atlantic Ocean in 1972 with the Chow Hall just next door - so I got by, and they were worth every penny!

When searching for a crossover kit, I left out the word "crossover" entered: AR3a rebuild kit. Four items are listen, but only one has do with AR3a speakers. Sadly, the link just takes you to a book seller. The odd part is, it reports it found 14 results.

Inserting the word crossover brings up 1 result, the bookstore.

Revising the search to AR3a crossover kit brings up 4 results but 3 go to the same site - and it says there are 18 results.

Same searches in Google yields less detailed results, but 100s more useful results, including crossover parts lists, schematics, suppliers, clubs, and more for these vintage audio history makers. cuil just isn't very kewl!

rokytnji · #19 (**permalink**) 21-08-2008, 04:39 PM

Haven't had time yet to test these search engines out yet. I have got into the habit of booking new ones as I find out about them. Some are google affiliated but minimal some not.

Ixquick Metasearch

goosh.org - the unofficial google shell.

Blackle - Energy Saving Search

Dogpile UK Web Search Home Page

Doogle

Just thought I'd throw these in the mix.

k_9 · #20 (**permalink**) 22-08-2008, 09:44 AM

Lol, Doogle made me laugh! But why would they want a link for Eurovision, they'll have to wait another year to redeem themselves. Gosh, i hope that one wasn't broadcasted in the media, imagine what remarks they'd get! Well, their search engine will sure die out. What we need here is an Irish guy to shed some light on us.

Now... who the hell uses Doogle? I've never heard of it before. However, I see that anything to do with leprechans or gobsh!te turns up on my favourites.