RE: [xml-dev] Discover data patterns or Create data patterns?

["Len" <cbullard@hiwaay.net>]

Yes, we have technology for making these systems, but Roger is asking a
subtler question:   How are discovery and creation different?

Patterns are discovered.   Sets of named patterns can be created.

How do we name names?

How is the problem domain named?   

Who picks the problem?

Discoverer or preparer?  

Are you creating a pattern by extracting and naming patterns from an
instance or are you discovering the pattern by which the extraction patterns
are selected?   

Is the selection of the name pre or post processed, in other words, named
selectors or selectors being named?

Are where clauses discovered?

Elements or attributes first?

Naming authority is political except by discovery because names are shared
among users and systems.   Discovery in an instance is local except where
the instance has referents and then only where the local instance is still
being discovered (update/referent tension (database ownerwship)).   

As Michael points out, the type can be changed for the value set by local
authority except where the value set has referents.

len


From: Mukul Gandhi [mailto:gandhi.mukul@gmail.com] 

I think, discovering data patterns is about studying the problem
domain, and preparing some kind of ER model (a logical data model) to
start with. We can then prepare a physical data model from a logical
model. The physical data model is targeted for a RDBMS schema, a XSD
schema or whatever storage we are interested in.