W3C home > Mailing lists > Public > public-lod@w3.org > November 2010

Re: Google Refine 2.0

From: Kingsley Idehen <kidehen@openlinksw.com>
Date: Fri, 12 Nov 2010 09:52:07 -0500
Message-ID: <4CDD5497.4040500@openlinksw.com>
To: Leigh Dodds <leigh.dodds@talis.com>
CC: David Huynh <dfhuynh@google.com>, Linked Data community <public-lod@w3.org>
On 11/12/10 8:35 AM, Leigh Dodds wrote:
> Hi Kingsley:
>
> I recommend you take some time to work with Refine, watch the demos,
> and perhaps read the paper that Richard et al published on how they
> have used and extended Refine (or Gridworks as it was)
Leigh,

You are repeating David's comments to me, which I deliberately ignored 
re. undertone.

Please Google up on pattern: gridworks kidehen. I celebrated the 
innovation before most, David and I do communicate offline too, 
typically he pings me when he has something exciting, and we have our 
little back and forth about issues we've argued about since 2007.

I am not saying I don't know what Google Refine does. I am not alien to 
data reconciliation, I am curious about the end game i.e., options for 
put the data in other data spaces beyond Freebase.

David:
I hope you understand that if a response doesn't start with: 
"Congratulations David...", it doesn't mean I am criticizing your work. 
You know me much better than that, I hope, as per my comments above to 
Leigh.

I just asked a question, where the focus of the question was scoped to 
an area of Google Refine that I hadn't looked into i.e., beyond its core 
ETL functionality. Again, an aspect, not the whole thing.

FWIW - I watched the video after sending my initial mail, and it didn't 
answer my question re. endgame. None of that diminishes the splendor of 
Google Refine. Anyway, when we're done with Pivot, a lot of the 
ramblings we had (offline) should become much clearer i.e., the area 
that I've always been interested in i.e., making Linked Data absolute 
fun for end-users, and in the process evolve them into "Citizen Data 
Analysts".  We took this journey once before via ODBC, but ODBC has 
platform specificity, data model, and data representation limitations 
that don't exist in the Linked Data realm. On the other hand though, 
ODBC ecosystem established solid patterns (loose coupling of compliant 
applications and drivers) that made it fun -- once you got past the 
aforementioned shortcomings.

> But to answer you question:
>
> On 12 November 2010 13:23, Kingsley Idehen<kidehen@openlinksw.com>  wrote:
>> How does the DERI effort differ from yours, if at all?
> They have produced a plugin that complements the ability to map a
> table structure to a Freebase schema and graph, by providing the same
> functionality for RDF. So a simple way to define how RDF should be
> generated from data in a Refine project, using either existing or
> custom schemas.

Thanks for the answer which means: Yes, to the issue of an RDF output 
option. But, unclear re. writing data directly to a SPARQL compliant 
Quad / Triple store.

> The end result can then be exported using various serialisations.
Naturally, once its RDF.
> My extension simply extends that further by providing the ability to
> POST the data to a Talis Platform store.

Yes, but why not any SPARUL endpoint since we have a standard in place? 
Which means the LODCloud benefits re. data quality etc?

>   It'd be trivial to tweak that
> code to support POSTing to another resource, or wrapping the data into
> a SPARUL insert

Yes, so why not make the tweak. ASAP?
> Ideally it'd be nice to roll the core of this into the DERI extension
> for wider use.

Good idea re. LATC project :-)

> Cheers,
>
> L.


-- 

Regards,

Kingsley Idehen	
President&  CEO
OpenLink Software
Web: http://www.openlinksw.com
Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca: kidehen
Received on Friday, 12 November 2010 14:52:40 UTC

This archive was generated by hypermail 2.4.0 : Thursday, 24 March 2022 20:29:51 UTC