On 4/13/14 8:32 PM, Patel-Schneider, Peter wrote:
On Apr 11, 2014, at 11:58 AM, Kingsley Idehen <kide...@openlinksw.com> wrote:On 4/11/14 2:12 PM, Patel-Schneider, Peter wrote:This proposal illustrates one of the major problems with the DBpedia ontology - triples check in but they never check out.Sorta, because of the misconception that SPARQL is steal Read-Only. I spend a good chunk of my day writing SPARQL 1.1 using INSERT, DELETE, and UPDATE (via INSERT and DELETE combos) to massage data, across many data spaces.How can I use SPARQL 1.1 to change the DBpedia ontology?
You can use SPARQL 1.1 (from your SPARQL 1.1 compliant application) to generate triples is a named graph local to your application, based solutions returned to you from the public SPARQL endpoint.
You can use LOAD, INSERT etc.. to produce your local RDF statements expressing whatever you have in mind. Once done, you can publish an RDF document for incorporation back into the DBpedia project etc..
Many of the problems with the DBpedia ontology require changes.Change is good. And it will work. Remember, DBpedia deploys Linked Data using a Quad Store, so the re-write rules and SPARQL queries used to perform the name->address indirection are extremely flexible. For instance, you can actually set the named graph URI scope for these SPARQL queries explicitly or via our NOT FROM NAMED GRAPH extension re., negation etc.. You can have a list of Named Graphs or excluded Named Graphs when generating the description of a DBpedia Entity URI's referent.I'm not sure how any of this can be used to effect changes in the DBpedia ontology.
Until you make a local copy, as I described above, you will believe the statement to be true. All you are doing (ultimately) is make changes in a document and then sending them over for incorporation. In the very worst case, you RDF document content will still be part of the LOD Cloud as long as you publish it on the Web.
The first step is making an RDF document with the alternative view that you seek.
Other problems have to do with the expressivity of the ontology.Again, its just triples to which SPARQL 1.1 patterns can be applied. In short, this is the way to truly appreciate the power of SPARQL and Linked Data.Again, I'm not sure what play SPARQL 1.1 has with respect to the expressive power of the ontology.
SPARQL 1.1 let's you make new RDF statements from existing RDF statements. The DBpedia ontology is a collection of RDF statements.
I don't think that changes here can be effected just by adding and removing bits of the ontology.You can variants of the Ontology, an alternative Ontology, it doesn't matter, the Linked Data deployment will be unaffected.Well, sure, none of this will affect most Linked Data uses, but that's not what I'm interested in. I'm interested in using the DBpedia ontology to organize information.
Yes, and information lives in documents, right? Thus, you simply grab the relevant data from DBpedia, massage it (using SPARQL 1.1 or other means) and then you have a new document (comprised of new or revised RDF statements).
I could, of course, simply use a different ontology, but my hope here is that use of the DBpedia ontology in products will result in a better ontology, and that that can be shared.
Yes, why is contributing tweaks to the DBpedia ontology using an RDF document produced by you not an option here? Do that, and everything else falls into place.
Other problems have to do with the philosophy of the ontology.The philosophy of the Ontology cannot change, that's a world view of the ontology creators. That doesn't stop another ontology existing as an alternative set of "context lenses" into the same data.It appears that the current ontology does not match the stated philosophy of the ontology. One or the other should change, and probably both.
Can't you reflect that in an RDF document submitted to the project?
I encourage you to make your changes, or make a new ontology, whichever path you take, the end product will be useful and a showcase for perspectives sometimes overlooked due to blurred and blurry perspectives :-)I would love to make changes. There actually is a modified version of the ontology that is in use.
Is this ontology represented in an RDF document that's accessible via an HTTP URL, at this point in time? If it exists, then we are nearly there.
Kingsley
peterWe can do this, its the next stage in the natural evolution of DBpedia and the broader Linked Open Data Cloud. Note: there is zero speculation in my response. I've already done (and continue to do) a lot of this (hands on fashion) over the years, following LOD cloud initial bootstrap. BTW -- This mail was easy to write as I've just completed (literally minutes ago) a marathon session on SPARQL 1.1 which amounted to an ACL (data access policy and access control list) debugging session. Kingsley.
-- Regards, Kingsley Idehen Founder & CEO OpenLink Software Company Web: http://www.openlinksw.com Personal Weblog: http://www.openlinksw.com/blog/~kidehen Twitter Profile: https://twitter.com/kidehen Google+ Profile: https://plus.google.com/+KingsleyIdehen/about LinkedIn Profile: http://www.linkedin.com/in/kidehen
smime.p7s
Description: S/MIME Cryptographic Signature
------------------------------------------------------------------------------ Learn Graph Databases - Download FREE O'Reilly Book "Graph Databases" is the definitive new guide to graph databases and their applications. Written by three acclaimed leaders in the field, this first edition is now available. Download your free book today! http://p.sf.net/sfu/NeoTech
_______________________________________________ Dbpedia-discussion mailing list Dbpedia-discussion@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dbpedia-discussion