On 9/8/10 9:16 AM, John F. Sowa wrote:
> On 9/8/2010 8:29 AM, Kingsley Idehen wrote:
>> I agree with most of your comments above, but we need to be a little
>> clearer about Linked Data (HTTP URI based hyperdata graphs) and Linked
>> Open Data (LOD) which is an application of the Web of Linked Data
>> publishing practices espoused in TimBL's meme [1].
> I'm glad that we mutually mostly agree.
>
>> LOD was bootstrapped through DBpedia [3]. This particular project
>> deliberately demoted RDF/XML to a negotiated data serialization format
>> while producing data sets culled from Wikipedia extraction and
>> transformation using N-Triples and N3/Turtle as preferred markup. Most
>> important of all, even before RDFa was incorporated in DBpedia, it
>> introducted the notion of HTML based entity descriptors, enabling easier
>> human readability and exploration of hyperdata graphs via Web Browsers.
> That is an excellent short summary that illustrates the basic issues.
> I would just add three points:
>
> 1. DBpedia is a community effort, and I'm all in favor of community
> efforts.
>
> 2. But I also believe that NLP technology has reached a point where
> automated and semi-automated methods can supplement, extend, and
> facilitate the development of such resources. (01)
John, (02)
Absolutely! This is basically what we do with tools like OpenCalais,
Zemanta, and Alcyhemy API. Basically, we have a piece of "on the fly"
RDF transformation middleware that we call the "sponger" [1][2][3] that
provides hooks for any NLP service that exposes an API. It's
architectural components are as follows: (03)
1. Extractor Cartridge (or Driver) -- transforms content into RDF and/or
enriches RDF it encounters (04)
2. Meta Cartridge (or Driver) - enriches output from #1 by passing text
through NLP engines and then performing LOD cloud lookups en route to
final Linked Data graph construction.
> 3. The automated tools will still require human guidance, review,
> and oversight, but they can be faster, more consistent, and
> more thorough than traditional methods. (05)
Yes, this is why annotation tools are so vital to the emerging Web of
Linked Data. Ironically, we don't actually many tools addressing this
critical area at the current time. Basically, the notion of a "Data
Wiki" is in plain sight, but only a few (including TimBL) have picked up
on this. TimBL's solution is that its implemented via the Tabulator
which is browser specific (re. extension) and so-so to use in hosted
form. We have solutions too, in many forms that inclued: QBE styley
Query Builders [4][5] with SPARUL capability, Annotation features within
our ODS [6] product etc.. But again, these products have questionable UI
(User Interaction) patterns (in the eyes of their beholders) due the
broader underlying concepts of Linked Data not being fully assimilated
by end-users, at the current time. (06)
Links: (07)
1. http://virtuoso.openlinksw.com/dataspace/dav/wiki/Main/VirtSponger --
Virtuoso Sponger Middleware
2.
http://virtuoso.openlinksw.com/screencasts/virtuoso-rdf-middleware3.swf
-- Virtuoso Sponger Workflow Animation
3. http://uriburner.com -- Sponger Instance behind a domain that
delivers RDF transformation and Linked Data generation & publication
service
4.
http://wikis.openlinksw.com/dataspace/owiki/wiki/OATWikiWeb/InteractiveSparqlQueryBuilder (08)
-- iSPARQL Query Builder
5.
http://wikis.openlinksw.com/dataspace/owiki/wiki/OATWikiWeb/InteractiveSparqlQueryBuilderTutorials (09)
- iSPARQL Query Builder Tutorials
6. http://ods.openlinksw.com/dataspace/dav/wiki/ODS -- OpenLink Data
Spaces (ODS) . (010)
Kingsley (011)
>> Yes! There is really absolutely nothing new about RDF or Linked Data.
>> Hopefully, people will soon come to understand this :-)
>>
>> Links:
>>
>> 1. http://www.w3.org/DesignIssues/LinkedData.html - TimBL's how to
>> publish Linked Data meme
>> 2. http://esw.w3.org/SweoIG/TaskForces/CommunityProjects/LinkingOpenData
>> -- Linked Open Data Community Portal
>> 3. http://dbpedia.org .
> Amen.
>
> John
>
> _________________________________________________________________
> Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
> Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
> Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
> Shared Files: http://ontolog.cim3.net/file/
> Community Wiki: http://ontolog.cim3.net/wiki/
> To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
> To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx
>
> (012)
-- (013)
Regards, (014)
Kingsley Idehen
President& CEO
OpenLink Software
Web: http://www.openlinksw.com
Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca: kidehen (015)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (016)
|