utypes: a proposal

[Douglas Tody]

On Fri, 31 Oct 2008, Igor Chilingarian wrote:

> I don't understand why not to follow this idea. Presently, UTYPEs look like 
> "almost-XPath" but not exactly. The differences are: (1) "." instead of "/"; 
> (2) no "[]" with additional constraints; (3) no reference to a namespace 
> (this seems to change with the latest Francois' proposal).

I disagree; aside from being inconsistent with current (standardized,
published, implemented, community invested) usage, basically I think
this is an attempt to replace UTYPE with what is essentially XPath.

UTYPE is part of the data model abstraction mechanism, intended
to separate the data model abstraction from implementation, a
fundamentally important concept where data models are concerned.
XPath is a technology-dependent implementation.  It is a runtime
expression, not a static type reference.

If we really need this, lets call it a UFI and deal with it as a
separate runtime mechanism for navigating complex data instances.
Lets keep UTYPEs simple.  UTYPEs are in the user interface.
Complexity kills projects like VO.

 	- Doug