Re: Tag ontology RFC

Danny Ayers wrote:
> On Apr 6, 2005 6:56 PM, Stefano Mazzocchi <stefanom@mit.edu> wrote:
> 
> Hi Stefano, glad you're chipping in. 

My pleasure if I can be helpful.

> I need to think about your other
> points some more, but I've an immediate response to this bit:
> 
> 
>>The idea of 'grouping-via-namespace' is very XMLish and not sure applies
>>very well here, 
> 
> Grouping *names* by namespace is a pretty fundamental idea,
> irrespective of how XML/RDF does it.

Sure, "grouping" is fundamental and the whole point of categorization is 
leading toward a form of clustering, but I'm not sure that using XML 
namespaces is the best way to achieve that in the folksologies space.

> I think it applies well here as
> it provides an easy way of creating tag lexicons, take these sets:
> 
> The del.icio.us "community" tags:
> 
> http://del.icio.us/tag/semantic
> http://del.icio.us/tag/web
> ...
> 
> The community tags as used by me:
> 
> http://del.icio.us/danja/semantic
> http://del.icio.us/danja/web
> ...
> 
> Now at present, I haven't got terms defined anywhere in my own
> namespace/domain corresponding to these, although I do have:
> 
> http://dannyayers.com/archives/author/site-admin/skos.rdf#c2
> 
> which is the intersection of the meaning associated with the terms
> "semantic" and "web" as used by me (which are equivalent to
> http://del.icio.us/danja/semantic and http://del.icio.us/danja/web),
> and there are a bunch of resource instances which are found carrying
> that classification:
> 
> http://dannyayers.com/archives/category/virtual-world/semantic-web
> 
> All the names above are HTTP-gettable URIs, but the del.icio.us style
> results in a much more convenient syntax.
> 
> Digressing a little, (heh, before I've thought about your points), the
> logical and human semantics are obviously open for discussion, but I
> do think there's an even simpler pragmatic interpretation available -

As I wrote in my blog post, I think that 'semantic' URI are going to 
harm you later on. But I also agree that this is just a subjective matter.

> Take a blog item, http://example.org/blog/blah. It turns up in the RSS
> feed (abbreviated) as:
> 
> http://example.org/blog/blah rdf:type rss:item
> 
> now Morten's FOAF Output is currently providing the following kind of
> thing for my named categories:
> 
> http://example.org/blog/blah dc:subject xxx:conceptX
> and
> xxx:conceptX rdf:type skos:Concept
> 
> which is quite indirect. I'm wondering what the harm would be in saying: 
> 
> http://example.org/blog/blah rdf:type xxx:tagX
> and
> xxx:tagX rdf:type x:Tag
> (and
> xxx:tagX rdfs:label "tagx")
> 
> with the human interpretation that the blog item is in the class of
> things tagged with "tagX".
> 
> but at the end of the day, what's important for
> 
>>interoperability is that the identifiers for your tags are globally
>>unique, the rest is just personal taste.
> 
> 
> Indeed.

Glad to find agreement here: I wish the handle/DOI/LSDI people would 
recognize that as well ;-)

-- 
Stefano Mazzocchi
Research Scientist                 Digital Libraries Research Group
Massachusetts Institute of Technology            location: E25-131C
77 Massachusetts Ave                   telephone: +1 (617) 253-1096
Cambridge, MA  02139-4307              email: stefanom at mit . edu
-------------------------------------------------------------------

Received on Thursday, 7 April 2005 01:32:01 UTC