Simon and I have been working on two Enterprise Ontologies lately. While we’ve both been involved in both of them, he took the lead in reducing what we’d learned to the model for one, and I the other. 

 

Without prior agreement, we each took different routes to the process first design. I rounded up all the properties that had come up in the course of over a dozen interactive elicitation sessions, and decided to start there. I thought if I could rationalize them and reduce their number, surely that would be a route to elegance.

 

 

Simon decided to rationalize them as he went, so he too had the product of about a dozen sessions, but decided to deal with the properties as he came to them.

I was quite chagrinned to find that even with all my rationalization I had twice as many properties as he did, and I still had to go back and prune those that didn’t get used. 

 

To be fair to the approach, Simon was able to out economize me in his model because he relied even more heavily than I did on properties that already existed in gist (our upper ontology). 

 

So while I think it still makes sense to focus on the properties, we can create an even more elegant model by reusing properties even more than you would normally expect.

Simon and I have been working on two Enterprise Ontologies lately. While we’ve both been involved in both of them, he took the lead in reducing what we’d learned to the model for one, and I the other. 

 

Without prior agreement, we each took different routes to the process first design. I rounded up all the properties that had come up in the course of over a dozen interactive elicitation sessions, and decided to start there. I thought if I could rationalize them and reduce their number, surely that would be a route to elegance.

 

Simon decided to rationalize them as he went, so he too had the product of about a dozen sessions, but decided to deal with the properties as he came to them.

I was chagrined to find that even with all my rationalization I had twice as many properties as he did, and I still had to go back and prune those that didn’t get used. 

 

To be fair to the approach, Simon was able to out-economize me in his model because he relied even more heavily than I did on properties that already existed in gist (our upper ontology). 

 

So while I think it still makes sense to focus on the properties, we can create an even more elegant model by reusing properties even more than you would normally expect.