Comments on: A Re-Introduction to UMBEL http://www.mkbergman.com/430/a-re-introduction-to-umbel/ Mike Bergman on the semantic Web and structured Web Mon, 06 Feb 2012 00:15:33 +0000 hourly 1 http://wordpress.org/?v=3.2.1 By: Exploding the Domain: UMBEL Web Services by Zitgist at Frederick Giasson’s Weblog http://www.mkbergman.com/430/a-re-introduction-to-umbel/comment-page-1/#comment-48102 Exploding the Domain: UMBEL Web Services by Zitgist at Frederick Giasson’s Weblog Sun, 20 Apr 2008 14:41:05 +0000 http://www.mkbergman.com/?p=430#comment-48102 [...] Part 1: A Re-Introduction to UMBEL [...] [...] Part 1: A Re-Introduction to UMBEL [...]

]]> By: The Cyc Foundation » UMBEL http://www.mkbergman.com/430/a-re-introduction-to-umbel/comment-page-1/#comment-47879 The Cyc Foundation » UMBEL Mon, 31 Mar 2008 14:32:41 +0000 http://www.mkbergman.com/?p=430#comment-47879 [...] Bergman began an interesting series about the history and motivations of the UMBEL project over the weekend. Stay [...] [...] Bergman began an interesting series about the history and motivations of the UMBEL project over the weekend. Stay [...]

]]> By: fred http://www.mkbergman.com/430/a-re-introduction-to-umbel/comment-page-1/#comment-47878 fred Mon, 31 Mar 2008 13:54:22 +0000 http://www.mkbergman.com/?p=430#comment-47878 Hi Danny, The question is: the hub of what? DBPedia/Wikipedia are mainly describing named entities, so individuals. In fact, in LOD, the only property that links "documents" and "papers" is owl:sameAs. As you know, this properly only says that two "Individuals" are the same. But what does this really say about the state of LOD? It says that currently, LOD only "links" the individuals of classes; and this is clearly the case with DBpedia. UMBEL goes one step further: it links classes of individuals. As this will be explained in the second part that Mike will publish tomorrow, UMBEL is also about how to CLASSify the linked data spaces. And our third article tries is to answer the question: why did we use Cyc as the reference framework? Some will like it, others won't. In fact, we welcome alternatives arguing that ANY form of context is better than none. I can already think about comments such as: why this and this? What is the utility? DBPedia already is linked to OpenCyc and has classes with categories from Wikipedia... and such. But believe me, we didn't put all that work on this without knowing these things, and there are reasons why we continued in this direction and why we will release this work in the next weeks and months. Take care, Fred Hi Danny,

The question is: the hub of what? DBPedia/Wikipedia are mainly describing named entities, so individuals. In fact, in LOD, the only property that links “documents” and “papers” is owl:sameAs. As you know, this properly only says that two “Individuals” are the same. But what does this really say about the state of LOD? It says that currently, LOD only “links” the individuals of classes; and this is clearly the case with DBpedia.

UMBEL goes one step further: it links classes of individuals. As this will be explained in the second part that Mike will publish tomorrow, UMBEL is also about how to CLASSify the linked data spaces.

And our third article tries is to answer the question: why did we use Cyc as the reference framework?

Some will like it, others won’t. In fact, we welcome alternatives arguing that ANY form of context is better than none.

I can already think about comments such as: why this and this? What is the utility? DBPedia already is linked to OpenCyc and has classes with categories from Wikipedia… and such. But believe me, we didn’t put all that work on this without knowing these things, and there are reasons why we continued in this direction and why we will release this work in the next weeks and months.

Take care,

Fred

]]> By: Danny http://www.mkbergman.com/430/a-re-introduction-to-umbel/comment-page-1/#comment-47872 Danny Mon, 31 Mar 2008 12:01:46 +0000 http://www.mkbergman.com/?p=430#comment-47872 While I'm far from convinced by the upper ontology approach - it's not very webby - something similar (but less philosophical!) could be very useful. Anything that stands a chance of increasing data linkage is worth a try. Right now DBpedia/Wikipedia seems to be the hub, so if I were you I'd make connecting up with that a priority. Like DBpedia/Wikipedia I reckon you've jumped a big hurdle with "a premise of emphasizing representational concepts over unattainable precision or exactitude". Couple of random refs you probably already have: <a href="http://suo.ieee.org/" rel="nofollow">SUO WG</a> , Sowa's <a href="http://www.jfsowa.com/ontology/toplevel.htm" rel="nofollow">Top-Level Categories</a> - his book's very good but hard work. While I’m far from convinced by the upper ontology approach – it’s not very webby – something similar (but less philosophical!) could be very useful. Anything that stands a chance of increasing data linkage is worth a try. Right now DBpedia/Wikipedia seems to be the hub, so if I were you I’d make connecting up with that a priority. Like DBpedia/Wikipedia I reckon you’ve jumped a big hurdle with “a premise of emphasizing representational concepts over unattainable precision or exactitude”.

Couple of random refs you probably already have: SUO WG , Sowa’s Top-Level Categories – his book’s very good but hard work.

]]>