[ExI] Google has incorporated Wolfram Alpha search results

Alejandro Dubrovsky alito at organicrobot.com
Sun May 24 10:30:16 UTC 2009


On Sun, 2009-05-24 at 10:01 +0000, BillK wrote:
> On 5/24/09, Alejandro Dubrovsky wrote:
> > I'm pretty sure they don't just feed the query you just put in to
> >  Wolfram.   That would never work in real time, especially because Google
> >  is probably about 20 times faster than Wolfram at responding. They are
> >  likely only giving you back the queries that were popular that had
> >  something useful on Wolfram, or it might even be Wolfram feeding this to
> >  the Google crawler through the robot.txt file.
> >
> >  eg try:
> >  x^2 +sin(x)+1284*cos(x)
> >
> >  You won't get anything back about Wolfram in Google (I assume. I tried
> >  something similar just before, but I don't want to pollute the cache
> >  with the exact query).
> 
> 
> After more playing, you are correct. Google isn't doing a Wolfram
> search. But Google must be finding that Wolfram result from somewhere.
> Even other similar simple equations do not produce any Wolfram
> results. Maybe that particular equation is a special Wolfram example.
> It is in the Google cache from 21st May, but it does link to a real
> Wolfram page.
> 
Google did do a Wolfram search at some point, just not at the time when
you searched for the term in Google.  It searched it before and has the
results cached.  A search in Wolfram is no different than a visit to a
page.  Google does this for lots of sites.  

It's hard to tell what triggered Google to perform that specific search
in Wolfram.  Most likely, it was just a link on a page  (ie on some page
someone wrote 'check this out, wolfram can do graphs for x^2 + sin(x)  
<http://www.wolframalpha.com/input/?i=x^2%2Bsin(x)> '
)
crawler followed, and the results got cached.
But it also could be wolfram's robot.txt file telling google to look for
that, or it could be Google picking up on this as a popular search and
telling its crawler to look at wolfram's because it looks like the kind
of thing that wolfram would know about.  This last option would have
taken a decision by someone in the company (not on those specific
searches, but on the idea of some bunch of queries being related to
wolfram), but the other two would happen naturally irrespective of the
target being wolfram, imdb or whatever.






More information about the extropy-chat mailing list