http://davedolan.com/blog/2007/04/10/the-associative-model-of-knowledge/ Website not included. Fri, 21 Nov 2008 04:37:26 +0000 http://wordpress.org/?v=2.6.3 http://davedolan.com/blog/2007/04/10/the-associative-model-of-knowledge/#comment-14266 Dave Fri, 18 Jan 2008 20:10:32 +0000 http://davedolan.com/blog/2007/04/10/the-associative-model-of-knowledge/#comment-14266 True, it's not the only one, per se, but it's the only one that works the way that it does. For example in XML based representations, identity is a lookup game. You have some attribute that you have to specify at some point in the document as an identity, and then reference it somewhere else, and in order to get that reference you have to traverse something, like an index if you're lucky, or the whole document if you're not. The uniquely vector addressable nature of items and their associates is a huge advantage because there is a notion of continuity inherent in both the conceptual and physical representations of it in the system. It's nearly a one to one mapping of reality to a model with very little impedance mismatch. I'm not discounting OWL wholesale of course there are a ton of things that it's great for, as well as relational databases, but as for economics of scale, the associative model (both semantic web and relavance) has a large potential market itch that relational datawarehousing has not yet begun to be able to scratch. The examples I've presented here are of course trivial, but if you get them, you can see how it can scale outward. More details to come shortly from me, I'm writing another monstrous post on the new system that's about to release any day now. True, it’s not the only one, per se, but it’s the only one that works the way that it does. For example in XML based representations, identity is a lookup game. You have some attribute that you have to specify at some point in the document as an identity, and then reference it somewhere else, and in order to get that reference you have to traverse something, like an index if you’re lucky, or the whole document if you’re not. The uniquely vector addressable nature of items and their associates is a huge advantage because there is a notion of continuity inherent in both the conceptual and physical representations of it in the system. It’s nearly a one to one mapping of reality to a model with very little impedance mismatch. I’m not discounting OWL wholesale of course there are a ton of things that it’s great for, as well as relational databases, but as for economics of scale, the associative model (both semantic web and relavance) has a large potential market itch that relational datawarehousing has not yet begun to be able to scratch. The examples I’ve presented here are of course trivial, but if you get them, you can see how it can scale outward. More details to come shortly from me, I’m writing another monstrous post on the new system that’s about to release any day now.

]]> http://davedolan.com/blog/2007/04/10/the-associative-model-of-knowledge/#comment-14265 Darroll Fri, 18 Jan 2008 20:03:19 +0000 http://davedolan.com/blog/2007/04/10/the-associative-model-of-knowledge/#comment-14265 Relevance isn't the only ones using Associative Databases. The semantic web is founded on an associative model. I'm working on an in house associative database written in Oracle and C#. As an early adopter I can tell you that there is still a lot of things to figure out for this model to compete against relational databases. Relevance isn’t the only ones using Associative Databases. The semantic web is founded on an associative model. I’m working on an in house associative database written in Oracle and C#. As an early adopter I can tell you that there is still a lot of things to figure out for this model to compete against relational databases.

]]>