On 11/18/11 5:24 AM, Steffen Lohmann wrote:
On 17.11.2011 20:03, Richard Cyganiak wrote:
Hi Steffen,

On 17 Nov 2011, at 14:34, Steffen Lohmann wrote:
MUTO should thus not be considered as yet another tagging ontology but as a unification of existing approaches.
I'm curious why you decided not to include mappings (equivalentClass, subProperty etc) to the existing approaches.

Good point, Richard. I thought about it but finally decided to separate these alignments from the core ontology - therefore the "MUTO Mappings Module" (http://muto.socialtagging.org/core/v1.html#Modules).

SIOC and SKOS can be nicely reused but aligning MUTO with the nine reviewed tagging ontologies is challenging and would result in a number of inconsistencies. This is mainly due to a different conceptual understanding of tagging and folksonomies in the various ontologies. To give some examples:

- Are tags with same labels merged in the ontology (i.e. are they one instance)?
- Is the number of tags per tagging limited to one or not?
- In case of semantic tagging: Are single tags or complete taggings disambiguated?
- How are the creators of taggings linked?
- Are tags from private taggings visible to other users or not?

Apart from that, I would have risk that MUTO is no longer OWL Lite/DL which I consider important for a tagging ontology (reasoning of folksonomies).

The current version of the MUTO Mappings Module provides alignments to Newman's popular TAGS ontology (mainly for compatibility reasons). Have a look at it and you'll get an idea of the difficulties in correctly aligning MUTO with existing tagging ontologies.

Best,
Steffen

Steffen,

Mappings loaded, some links demonstrating effects:

1. http://linkeddata.uriburner.com/describe/?url=http%3A%2F%2Fpurl.org%2Fmuto%2Fcore%23hasCreator -- subPropertyOf relation with sioc:has_creator

2. http://linkeddata.uriburner.com/describe/?url=http%3A%2F%2Frdfs.org%2Fsioc%2Fns%23has_creator -- SIOC ontology

3. http://linkeddata.uriburner.com/describe/?url=http%3A%2F%2Fwww.openlinksw.com%2Fschemas%2Fgoogleplus%23comment_actor -- Google+ ontology

4. http://linkeddata.uriburner.com/describe/?url=http%3A%2F%2Fwww.openlinksw.com%2Fschemas%2Fgoogleplus%23Actor -- G+ ontology instance data.

If I use your mapping ontology as basis for an inference rule, I won't need the steps above since the class instances will manifest for any classes in owl:equivalentClass relations. Same thing applies with our ontologies which map across many ontologies, basically the power of OWL inference enables one coral lots of instance data via inverse relations :-)


BTW - you can substitute linkeddata.uriburner.com with lod.openlinksw.com re., URLs above and perform similar exploration against the larger and more powerful LOD Cloud cache instance we maintain.

--

Regards,

Kingsley Idehen 
President&  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






Attachment: smime.p7s
Description: S/MIME Cryptographic Signature

Reply via email to