Ex-Googlers reinvent web search In an attempt to upstage their former employer, a trio of ex-Googlers have launched a search engine of their own. They insist on calling it "Cuil" - pronounced "cool," apparently - and they say it "goes beyond today’s search techniques." This includes flashing random pornography when you search for information about a Grenoble …
"Content based results" sounds fine in practice - but sometimes "most popular" really is what you want. This is particularly the case where the content is applied by javascript or an ajax call. Being able to handle THOSE kind of sites would be truly radical. (Overuse/misuse of ajax is to be avoided, but sometimes it's necessary...)
E.g. - arguably, searching on something like "diddlefinger" should bring up the google maps mashups by that name [disclaimer: it is mine] as one of the top links - not a range of sites that refer to it.
I searched on "Posix thread mutex" on cuil. It came back with links that seem well related to the search, but the pictures are a bit out there.
link title: "Boost.Threads - Mutex Concept", picture of book "Programming with POSIX Threads" by Butenhof.
link title: "Multithreading Programming Topics: Using POSIX Thread ...", picture of book "Programming with POSIX Threads" by Butenhof.
link title: "Qt Toolkit - Thread Support in Qt", picture of some cross stitch or petty point.
link title: "Thread Support in Qt", picture of a music CD "Star of Crash" (hard to read the title.
link title: "POSIX Threads Programming", picture of some men working on one of the cables atop a large suspension bridge.
I followed some of the links and they were indeed all related to POSIX threads, which is not too shabby. Just the pics were a bit odd.
Now, if I had searched on "POSIX rope mutex", or perhaps "insert object into STL container" I might have got some different pictures.
Interesting also that wikipedia did not constitute most of the first page of links.
Just tried searching for information on Cuil on Cuil.
Found Sligo and Scotland, but not Cuil the search engine!
If the search engine ignores self-promotion, it's got to be good!
[Paris, because searching for Paris on Cuil definitely gives more interesting results than searching for Paris on Google...]
I did a test search using House of Health and found that there were multiple duplications of a small number of sites and their contents, though there was a statement saying that "92,746 results for House of Health" only 23 pages with 11 results showed on each, most of the 264 were duplicates.
Google shows more House of Health sites.
Test for yourself.
I wrote them a grumpy email yesterday for using a copyrighted image that has nothing to do with me or or the content of my site to link to it. So it could have been worse, eh?
Also, the searching itself is complete pants. As mentioned above, some huge proportion of the search results are simply links to online directories (such as dmoz) - doing a search on "canberra organic garden" has my site on the first page... but no return on the Canberra Organic Gardening Society, half a dozen links to directory sites, and the balance having nothing to do with Australia, much less Canberra (or organic gardening). Searching on "organic garden canberra" doesn't return my site, nor COGS or anything else to do with organic gardening in Canberra. "organic garden_ing_ canberra" returns a whole raft of different results, with COGS finally showing up.
Similar results occur with all of the searches I've carried out, with occasional links to those Wikipedia clone sites when it comes to the more factual material.
Can I also point out that using a word to name your website that cannot be correctly pronounced by anyone other than Gaelic readers is complete w@nk. There are plenty of other languages that have words that can be pronounced by anyone who reads a Latin-based script (I vote for "kono", which is Maori for "basket" (of knowledge).
/Paris, because she can usually find what she's looking for, unlike Cuil.
...actually your use of ajax does not seem terribly extensive at all, and it certainly isn't the kind of site I had in mind. (It is a nice site, just not of the type i was referring to...)
No one doubts that you can often, with a bit of thought, "do things in a way that is search engine friendly" - but if that is the only reason that you are doing it- rather than because the application itself demands it, then you are compromising your site simply to get around an inadequacy in a search engine...
Pragmatically, that may be worth doing, but it is still more or less a "nasty hack". (That said, being forced to think "do I really need to do this on an ajax call or by writing to the DOM directly" is a jolly good discipline - but sometimes the answer is "yes". )
The point I was making is that going to a "content" model, as Cuil seems to, is something of a step backwards, because it exacerbates the problem with indexing content applied after the page has loaded - which it seems to be unable to parse. For such sites, the "popularity" model actually works better as web-surfers can analyze the content better than Cuil is apparently able to.
I think the UI has potential, but the results are rubbish. About half the searches I tried (all technical - honest!) returned name squats and really stupid pictures with no bearing on the subject matter.
Pictures only inform if they are relevant. If not, then leave them out.
Cuil claims to search more pages than Google. Perhaps, but quality is far more important than quantity.
"We make extensive use of Ajax without any problem whatsoever with search engines, eg. http://www.google.com/search?q=site:holyculturedownload.com&hl=en
This comes down very simply to superior site design."
Installing OSCommerce and throwing a template on it doesn't make you or your websites superior to anything.
"Maybe they just want to sell it to Microsoft for a big bag of money."
If there's astute Celtic blood coursing through their veins, then licensing it to Microsoft for bags of big money would have them sitting pretty and feeling on top of the world.
And with careful tweaking/tuning of the search algorithm, in order to deliver outstanding lead content, can easily ensure that Novel Information Leads Future Direction ..... which would be Real Cool Cuil ...... although paradoxically, it would be AIRed Hot XXXX dDevelopment ReSearch and Product Placement.
Are they Planning a Simply Cuil Cloud Complex to lay waste to mouldy G.C.H.E.E.S.E. and the Fuddy Duddy. As we know, cheese is one of those new fangled, entangled string fuels for Upper Crust Echelons ....... http://www.itwire.com/content/view/19179/1066/
Or is that something which is always in Locked and Secure Beta AIR&dD and therefore Intellectual Property which need not necessarily be discussed and/or revealed until such times as they have IT sussed or choose to XXXXPose.
Seems like a Good Plan ..... and for anyone capable of sitting on top of the world and feeling pretty ..... a Godsend ....... although Really, it is just IT and AI conspiring 42 EduTain you, Virtually....... with AIRange of Network InterNetworking RobotIQs to Capture the Living Soul and Process ITs Imagination into Beta CyberIntelAIgent Leads Transparently Shared Steganographically with Much SMARTer Leaderships.
Way to Go, Cuil. amfM HyperRadioProActivity Seconds that NEUKlearer Quantum Leap Path.
And yes, you can think that that is Absoluut GBIrish but any Full Monty Goon would Realise that it is also Raw PerlyGatesPython for Empirical Imperial Use.
cc HMGCC re GCHQ Roles in Great Games 42 Play with Quantum Communications BetaTesting Intelligences.
IT certainly beats pratting about like the headless chicken or psychotic hubristic idiot in Palaces like Westminster....... :-)
Just searched for the title of my own website, and 90% of the results were ad-sites with text from my website in the description. The description on Cuil, that is. Going to the site in question showed nothing of the sort on the page, just ads
The top result was a link to an old URL of my site, which has been off-line for over a year...
Thumbs down, cause they need to do some work before I step over from Scroogle...
"I wrote them a grumpy email yesterday for using a copyrighted image that has nothing to do with me or or the content of my site to link to it. So it could have been worse, eh?"
So did I - using a different charities logo for the charity website I run is not a good start (charities that work in the same kind of area, but different approaches and one secular, one not)
Not impressed so far...
Brilliant, I think it is a ploy, gosh I hope it is because if not I will be looking for vengeance.
Cuil it has an i not a double l :) But cool it is not.
I cannot even work out the correlation of picture to url logic, it is just one great cock up.
They could be road side bombers, have to check out some other urls (see that is the ploy).
But in a strangely logical way. Mixing up charity logos, for example. I searched for my band, and although it found our website (consisting of a single logo and a list of MP3s) it ignored our logo and presented the album cover from "A Looney Tunes Sing-A-Long Christmas". Which doesn't quite fit with our post-punk stylings, but is gratifying nonetheless.
Strangest thing is that it seems to be consistent - using a different wrong image each time would suggest a 'simple' threading issue. Using the *same* wrong image each time is just confusing (and prone to law suits).
Paris, because this comment has nothing to do with her, and if it's good enough for cuil...
Well let's see, the web is comprised of 35% junk (pics of pets or kids), 30% porn, 20% news, fake news and tech news, and the rest is either purely commercial or useless.
So if your search site wants me to discover the 70% I'm not interested in, don't bother.
Went to check it out - took about a minute to open, found nothing that Google didn't, presented the results in a cluttered mess and for "DHCP server Windows 2003" found no results. Best not to announce your "Google-beater" until you actually have it working. Have a Grolsch and get back to the drawing board.
A search for "Glasgow" returned a link for 'Glasgow Centre for the Child and Society' with an accompanying image of Osama bin Laden!
A search for "delicious" (or even "del.icio.us") failed to return the social bookmarking site on the first page of results.
It's not all bad though. "Reg" returns a top centre link to theregister.co.uk
But on balance: rubbish. I'll not be back in a hurry.
I find Scour (formerly aftervote, formerly younanimous) to be a much more useful alt search at the mo.
The naughty pic has now disappeared. Testing cuil a bit revealed it is still far from cool.
"i wanna rule the world" 10cc lyric returns 0 results on cuil, whilst removing the quotation marks returns 4 605 641 results as opposed to Google's 856. I did not check for duplications, though (I have to get some work done as well!).
Quotation marks, however, definitely confuses cuil no end, almost invariably returning the following error page:
We didn’t find any results for “"whatever search criteria you entered"”
Some reasons might be...
* a typo. Please check your spelling.
* your search includes a term that is very rare. Try to find a more common substitute.
* too many search terms. Please try fewer terms.
Finally, try to think of different words to describe your search.
@Anonymous Coward
"....'Cull' is pronounced 'cull' as in 'hull', if you want to call it 'cool' spell it 'cool', or 'cuul' or 'cewl'...."
that's all very interesting, but since the site is called 'cuil', not very relevant.
----
@gaelic scholars everywhere
'cuil' is irish for fly [as in buzzing insect] or 'bad mood' and is pronounced somewhere between 'quill' and 'kill' depending what part of the gaeltacht you're from. if they want it to be pronounced 'cool' they'd have to stick a 'fada' on the 'u' and spell it 'cúil'. that means a 'nook' or 'corner'.
i can see plenty of opportunity for witty quippage with any of those three options:
* 'fly' - because you're likely to land in the shit, using it
* 'bad mood' - because that's what you'll be in, when you search for yourself and find some gay porn under your name [unless you're a gay porn-star!]
* 'nook/corner' - because [judging on current performance] that's likely to be representative of the amount of google's market share this thing manages to steal
Totally bizarre, I think amanfrommars has been helping code their search algorithms...
Searching for my company name does find my homepage amongst lots of unrelated sites, but like others above they have used someone else's logo.
I note Cuil seem to be hosting the images themselves, so they maybe don't have enough storage / bandwidth and want to just re-use the same few files so they get cached by people's browsers or proxy servers to save on resending.
Results which link to my blog have pictures of some TV show (could be Lost as the bot probably was, but it's a very small thumbnail so hard to tell). Another link to my blog has a pic of a car showroom (nothing to do with me in any way whatsoever).
Even more weird, a search for "Microsoft training Leeds" has as the second result a page on www.tibetarts.com which does not exist, with link text from an Irish e-learning provider's press releases (nowhere near Yorskshire). Even more strangely, the title text carrying the link reads "management training lesbian training microsoft training".
This is just ridiculous, and has nothing to do with the linked site at all. Putting it back into cuil as a search term does not return that page at all.
The other results seem to have a preference for pictures of David Brent, for some unknown reason.
Does this non-existent page count as one of the 121 billion pages indexed?
I like the layout of the search results - it looks a bit bloggish, in a good way. Pictures will help the look but only they have at least a vague relevance. Not sure if they help you choose the right link to follow, though.
As for the name, I have no problem if the Gaelic for knowledge happens to sound like "kewl" - it's no more arbitrary than "Google", or misspelling words like digg and flickr.
The level of 'reliabiity' is laughable. The first test terms I tried worked perfectly, right site, No1 ranking, even the right image (the site banner) - after that though it went seriously down hill, failed to find any correct sites from test terms known to work on the big G within the first 2-3 pages. And tons of completely inappropriate, not to mention more than a few NSFW accompanying images.
In a way it reminds me a little of the early Google - but with that you did often find what you were after, the odd 'off the wall' result was a kind of unlooked for bonus.
This thing, except for the very first search I tried, isn't off the wall - it's missed the building, the town, the country and isn't even on the right planet!
Cool it definitely ain't. However you spell it.
I know, I know - why isn't there an edit button on this thing? I blame this French keyboard - I clearly intended to type a "4" there.
Although there is a boost, I think most people will discover that cuil is a bit rubbish.
I'm waiting for this to be added to the wikipedia page now.
I was trying my best to get NSFW stuff on my desktop. According to some AC, a search for "sql case" would do. It gives a tiny thumb, among others, and clicking on it, continuously ends here:
Server Error in '/' Application.
The remote server returned an error: (500) Internal Server Error.
Description: An unhandled exception occurred during the execution of the current web request. Please review the stack trace for more information about the error and where it originated in the code.
Exception Details: System.Net.WebException: The remote server returned an error: (500) Internal Server Error.
Source Error:
Line 48: Dim aRequestedHTML() as Byte
Line 49:
Line 50: aRequestedHTML = objWebClient.DownloadData(siteURL)
Line 51:
Line 52: 'STEP 3: Convert the Byte array into a String
Source File: c:\websites\cutebaskets293\cutebaskets.com\process.aspx Line: 50
Stack Trace:
[WebException: The remote server returned an error: (500) Internal Server Error.]
System.Net.HttpWebRequest.CheckFinalStatus() +680
System.Net.HttpWebRequest.EndGetResponse(IAsyncResult asyncResult) +139
System.Net.HttpWebRequest.GetResponse() +251
System.Net.WebClient.DownloadData(String address) +151
ASP.process_aspx.Page_Load(Object sender, EventArgs e) in c:\websites\cutebaskets293\cutebaskets.com\process.aspx:50
System.Web.UI.Control.OnLoad(EventArgs e) +67
System.Web.UI.Control.LoadRecursive() +35
System.Web.UI.Page.ProcessRequestMain() +750
Version Information: Microsoft .NET Framework Version:1.1.4322.2407; ASP.NET Version:1.1.4322.2407
That's Microsoft-dot-Net for you. No wonder, AND a confirmation for my worst suspicion: Of course, they never want anything with it! They make it somewhat workable, remove the pr0n, and sell it to Steve B. for a few hundred millions. Bastards!
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
