W3C home > Mailing lists > Public > public-lod@w3.org > May 2012

Re: Introducing the Knowledge Graph: things, not strings

From: Kingsley Idehen <kidehen@openlinksw.com>
Date: Sat, 19 May 2012 11:44:32 -0400
Message-ID: <4FB7BFE0.6090005@openlinksw.com>
To: public-lod@w3.org
On 5/17/12 11:55 AM, Joseph Montibello wrote:
> Hi all,
>
> Can someone interpret this for me? I took a google search that has an
> information box on the side that I think is what Google's referring to as
> their knowledge graph.
>
> http://bit.ly/googlelod
>
> I started looking at it using a chrome extension, OpenLink Data Explorer.
> I threw up a pdf of what I saw at
>
> http://bit.ly/googlelod2
>
> In that document, which lists out the triples found on the page, one of
> the triples is as follows:
>
> subject:
> search?hl=en&biw=1823&bih=850&q=jan+kerouac&stick=H4sIAAAAAAAAAONgVuLUz9U3M
> CmoSEoHAO-WRMoNAAAA&sa=X&ei=Vhm1T83QD6me6AGgg7DfDw&sqi=2&ved=0CLwBEJsTKAA
> predicate:     void:sparqlEndpoint
> object:           sparql
>
> Does anyone believe / can anyone confirm that this is an actual sparql
> endpoint?  Sorry if I'm totally crazy here - very new to looking at linked
> data in general, so I'm still feeling my way along.

That can't be right. At the very best, the triple should imply there is 
a sparql endpoint that provides access to the data transformed via the 
sponger linked data middleware. Thus, given a proxy (wrapper) URI, you 
can obtain its description (via describe or construct) via the sparql 
endpoint in question.

I assume you've looked at the examples at: http://ode.openlinksw.com or 
the notes at http://uriburner.com etc..

Kingsley
>
> Thanks,
> Joe Montibello, MLIS
> Library Systems Manager
> Dartmouth College Library
> 603.646.9394
> joseph.montibello@dartmouth.edu
>
>
>
>
>
>
> On 5/17/12 11:22 AM, "Steve Harris"<steve.harris@garlik.com>  wrote:
>
>> On 2012-05-17, at 00:04, Kingsley Idehen wrote:
>>
>>
>> On 5/16/12 6:55 PM, Bernard Vatant wrote:
>>
>> Adrian
>>
>> Don't dream of accessing the Google Knowledge Graph and query it through
>> a SPARQL endpoint as you do for DBpedia. As every Google critical
>> technological infrastructure, I'm afraid it will be well hidden under the
>> hood, and accessible only through the search
>> interface. If they ever expose the Graph objects through an API as they
>> do for Gmaps, now THAT would be really great news.
>>
>> Kingsley says they have Freebase, yes but Freebase stores only 22 million
>> entities according to their own stats, which makes less than 5% of the
>> overall figure, since Google claims 500 million nodes in the Knowledge
>> Graph, and growing.  So I guess they have
>> also DBpedia and VIAF and Geonames and you name it ... whatever open and
>> structured they can put their hands on. Linked data stuff whatever the
>> format.
>>
>> Bernard
>>
>>
>>
>> And it will be query accessible, this is something that's inevitable and
>> unavoidable. This is the Web.
>>
>>
>>
>>
>>
>> I doubt it. Google don't even allow API access to their search engine. I
>> can still remember the days when they were a search company ;)
>>
>>
>> For them it's all about staying ahead of the competition so they can get
>> more eyeballs on google ads, and more tracking data - interactions with
>> humans basically - providing APIs to their graph data doesn't help that
>> aim.
>>
>>
>> Doesn't mean they won't do it, but I don't think there's any reason for
>> them to.
>>
>>
>> - Steve
>>
>>
>> -- 
>> Steve Harris, CTO
>> Garlik, a part of Experian
>> 1-3 Halford Road, Richmond, TW10 6AW, UK
>> +44 20 8439 8203  http://www.garlik.com/
>> Registered in England and Wales 653331 VAT # 887 1335 93
>> Registered office: Landmark House, Experian Way, Nottingham, Notts, NG80
>> 1ZZ
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>
>
>
>


-- 

Regards,

Kingsley Idehen	
Founder&  CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca handle: @kidehen
Google+ Profile: https://plus.google.com/112399767740508618350/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen








Received on Saturday, 19 May 2012 15:44:57 UTC

This archive was generated by hypermail 2.3.1 : Sunday, 31 March 2013 14:24:39 UTC