[Top] [All Lists]

Re: [ontolog-forum] Semantic Web shortcomings [was Re:ANN: GoodRelations

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: Frank Guerino <Frank.Guerino@xxxxxxxxxxxxxx>
Date: Mon, 18 Aug 2008 11:47:56 -0400
Message-id: <C4CF11EC.17FC8%Frank.Guerino@xxxxxxxxxxxxxx>
Hi Duane,

I agree with this as a small piece of the overall puzzle.

For example,  one of many ways to do this is to instantly and dynamically synthesize advanced search forms from the ontological definition of data entities.  (I’m oversimplifying this, as you have to be able to do things like have the predefined static reference data to control pre-populated fields, which is significant, in and of itself.)

So, to continue your example, if someone searches for the string “washington”, it has no context.  However, if someone searches for the string “washington” in the last name field of a human or in the state field of a location or in the city field of a customer account, “washington” now has “context”.  Such context instantly works to filter out results that would have otherwise come back that had nothing to do with the intended search context.  This, I believe, correlates to your use of the description “The user could then disambiguate the search and refine their result set...

Doing all of this has a significant number of challenges that stand in the way.  The most important is the structure of the stored data.  Raw unstructured data, such as a MS Word document or flat file, typically cannot guarantee any real form of context for the data within it.  To facilitate context based search that leverages the ontology of a data entity, the data will typically have to be stored as a structured entity/noun.  This will allow values to be mapped to specific attribute names, attribute types, attribute constraints, etc., which is all hidden “context” about the data being stored.

Once you get all of this right, though, the results are pretty impressive.

Even so, solving the above problem then leads you to a whole new set of problems, such as searching through “relationships” that act as descriptive roles to get even more context, and so on.  This is how we get to Visual Mind-Maps, 360 Degree Graph Views, Tree Views, Heat Maps, etc. (i.e. the things that RDF and OWL are ultimately attempting to build a foundation for.)

Anyhow, I hope this adds value.

My Best,


On 8/16/08 1:12 PM, "Duane Nickull" <dnickull@xxxxxxxxx> wrote:

What I envision as a first step towards a true semantic web is a two step search interface, powered by ontological work, to disambiguate.
Users experience would be like this:

Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Subscribe/Config: 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 Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (01)

<Prev in Thread] Current Thread [Next in Thread>