Dump Google?

Michael Loftis mloftis at wgops.com
Sun Sep 12 16:16:08 UTC 2010

--On Saturday, September 11, 2010 10:05 PM -0700 Robert Holtzman 
<holtzm at cox.net> wrote:

> On Sat, Sep 11, 2010 at 05:34:25PM -0600, Michael Loftis wrote:
>> --On Saturday, September 11, 2010 3:33 PM -0700 Robert Holtzman
>> <holtzm at cox.net> wrote:
>> > On Sat, Sep 11, 2010 at 03:21:51PM -0400, Simon Ponder wrote:
>> >> What other engine do you use, if you do not mind me asking?
>> >
>> >            ..........snip..........
>> >
>> > Icerocket, although it's been getting flaky on me as of late.
>> Uhm, news flash.  Icerocket's web search IS GOOGLE.  I think it's blog
>> search is also google based, I'd have to dig, but, looks a bit like the
>> Google news or groups search.
> I did a little digging. Running a search on "icerocket + google" turned
> up several sites that contrasted icerocket and google. If there was
> anything linking the two, I missed it. Can you supply a URL for your
> conclusion?

The fact that their web search result pages are nearly identical to 
Google's (minus the upper header actually), and results are identical to 
Google.  Just do some comparison searches.  They find the same numbers of 
pages, rank them the same, and are using the same extracts/excerpts.

I really highly doubt they've enough spidering capacity to replicate 
Google's results so closely.  The fact that their nothing found/error page 
also contains Google's nothing found/error language verbatim points to this 
as well.

As for their blog search, it also looks like the Google Blog Search API 
Data, with some form of additional filtering, exactly what they're doing 
there I'm not sure.

Icerocket is very clearly someone whose written a UI for Google searches, 
there's nothing there to suggest otherwise.  In web searches especially 
they're *identical*.  The likelihood of two independent search databases of 
the web producing the EXACTLY same results for the first 15 for every 
single search I tested (I tried 6 of them, 'dog pile', 'google 
philanthropy', 'rock hunting', 'terranova space suit', 'feel good music', 
'hockey pucks for sale' -- just random keyword strings really except for 
the google philanthropy one).  And at a glance it also appears everything 
past the top 15 was identical too.  Empirically, Icerocket web search is 
just google search API.  If anyone here is self serving it's Icerocket.

Try matching ANY other search engine against Google, (or against any 
other!) You're not going to get the same results.  Even if they use the 
same algorithms, differing databases will produce different results.  The 
only way to replicate the breadth and depth of Google's results is to have 
the many many many TB of search index capability that Google has.

I'd be really surprised if their blog search isn't Google, the data that's 
there is what is represented in the API's.  That one I haven't been able to 
figure out what they're doing to get those results, so they're offering 
something of value there.  It certainly produces better results than 
blogsearch.google.com -- but maybe that's not the data stream that 
icerocket is using either.

The simple fact that they're blatantly lifting Google web search though 
makes it pretty likely their blog search is based off Google data.  The 
twitter search looks to me to be a wrapper around Twitter's own Search API 
as well, but I didn't spend any time looking into that.

Their 'advanced search' syntax, is also identical to Google's (that's not 
saying much honestly, but it's one additional little thing) -- though 
they're filtering out at least some of the specialty search prefixes like 

More information about the Ubuntu-devel-discuss mailing list