Re: Approaches to Expanding the Semantics of a Community's Self-Interest

[Marcus Carr <mcarr@allette.com.au>]

Costello, Roger L. wrote:

> I am documenting the different approaches for extending the semantics
> of a community's tag-set. I seek your thoughts on this topic.  

It seems to me that the characteristics of the communities and the data 
may mean that the choice is not straightforward. The residential lending 
chain in Australia uses a set of schemas to describe the information 
required to drive interactions such as broker/lender, lender/valuer, 
lender/mortgage insurer, etc. Assuming you consider these interactions 
to be representing communities, there is a very definite progression of 
various parts the same information through multiple communities. But the 
lending chain itself is also a community, consisting of the individual 
communities used to drive processes. A property valuer and a credit 
scoring agency never interact in the lending chain, but there are 
intersections of their data. Allowing them and every other community in 
the chain to represent the same address different ways is not ideal.

Your approach 1 or 2 might be better for the small communities, but they 
would may life very difficult for the chain, as change would ripple up 
and down. To eliminate that, we use a vocabulary that supports the 
superset of all structures, then use derivation by restriction to 
produce the final schemas for the various communities. Of course that 
means that we get people who suggest that there would be a cleaner way 
do describe data for their community and they're usually right. But the 
more flexibility we give them, the further away the realization of 
Straight Through Processing (STP) of a mortgage becomes. The key is to 
find the balance that allows them to describe all the necessary 
information, but to do it in a way consistent with the chain.


Marcus Carr