- From: Kjetil Kjernsmo <kjetil@kjernsmo.net>
- Date: Wed, 12 Sep 2012 14:06:13 +0200
- To: Enrique Pérez Arnaud <enriquepablo@gmail.com>
- Cc: semantic-web@w3.org
Onsdag 12 september 2012 01:22, skrev Enrique Pérez Arnaud: > I disagree :) traits, aspect orientation, behaviours, these are tools > designed to encourage code reuse in large frameworks, and they don't have > the inner structure that is needed in this case. In the same way as the > class system (OO) has the inner structure that is needed (and was also > designed for code reuse). The initial simplicity given by OO is not worth > the complication of injecting into it a more general deductive system. I > would say that the more general deduction system should be the core, and > then, you might add shortcuts into it, such as a class system. Hmmm, I wonder if we actually agree. :-) I'm not quite sure, though. Based on my earlier life as semweb software engineer, I had noted that developers, both with and without semweb background lacked tools in the toolchains to do stuff beyond simple generic browsing, and that the generation of classes (using e.g. a metaobject protocol implementation) at runtime would be a very interesting way to let them work with RDF data. Then, I came to realize that single inheritance is too inflexible and multiple inheritance, is, to be brief, too scary. My old example is: "say, you have a car and a boat, and reasonable implementations of those. Then an amphibious vehicle comes floating [sic!] by." The idea is that even though you have not anticipated the existance of such a thing when writing the Car and Boat classes, there exists most likely a composition of methods already implemented that would be useful and that you might be able to make that composition by using reasoning. I didn't intend to say that a general deductive system would be the right tool for the job (though I acknowledge that what I said might be interpreted that way), but that OWL reasoning might be able to do this for us. There has been some work that seem to have shared this vision, but I'm not aware of any research that have been fruitful. I admit that since my own research is very far from this, I have not been paying close attention, but what I'm saying is that it is a very interesting topic, and I'd be interested in references on the subject. Cheers, Kjetil -- Kjetil Kjernsmo PhD Research Fellow, University of Oslo, Norway Semantic Web / SPARQL Query Federation kjetil@kjernsmo.net http://www.kjetil.kjernsmo.net/
Received on Wednesday, 12 September 2012 12:06:44 UTC