Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

SUO: Latticework Ontology: The Virtues?




Hi,

This question is put primarily to John Sowa, but I'd welcome responses from 
anyone with a take on the topic.

I have tried to follow the entire debate on approaches, methodologies and 
in particular, the monolithic vs. latticework ontology issue. However, I'm 
not entire sure I've absorbed the salient points.

I find myself sympathetic (in a kind of non-specific way, alas) to the 
lattice advocates, since as a long-time programmer and software designer, I 
know that monolithic systems are problematic for a variety of reasons 
stemming directly from their monolithic nature.

Could I get a succinct, synoptic run-down on the purported advantages of 
the lattice-of-ontologies approach?

Here are a few things I can think might be the advantages:

- Modularity / Separability
- Composability
- Scalability
- Impact-of-change minimization
- Division of ontology authoring effort
- Accommodation of disparate formalizations
- Inference performance
- Inferential capabilities


A specific I question have about a lattice-structured ontology is whether a 
given symbol can be reused in conflicting ways in different 
non-overlapping, non-interacting component (sub-) ontologies?

Actually, I can probably think of other questions, but they're quite likely 
to be premised on my presumptions about the specific conceptualization of 
the lattice-structured ontology approach. I'll hold off enumerating them 
until I can get a better idea of why it is being promoted.

Thanks.

Randall Schulz
Mountain View, CA USA