Re: UDEF's day in the sun

[rschuldt <rschuldt@uswest.net>]

Robin,

Thank-you for the note to the group and for including the pointer from
your home page. As noted in past regrep meetings, I believe that a UDEF
based naming convention that conforms to ISO/IEC 11179 and that uses the
property terms (class words) of DoD 8320.1-M-1 is a necessary part of
the global registry and repository. The article by Bill Lewis does an
excellent job of highlighting the fact that XML needs an answer to the
semantic dispersion problem. Until a solution is widely adopted, the
various XML dialects will continue to generate applications that are
designed for computer-to-human-to-computer integration.

A joint AIA/EIA/DoD effort to create a National standard for Data
Interoperability and Configuration Management (EIA 836) is planning to
include the UDEF naming convention as part of the standard. The EIA 836
announcement is at:
http://www.geia.org/eoc/G33/g33index.htm

Ron Schuldt


Robin Cover wrote:

> Regrep member Ron Schuldt (rschuldt@uswest.net) has reminded us
> from time-to-time about the UDEF solution to the network's
> "semantic dispersion" problem.
>
> UDEF (The Universal Data Element Framework) has been identified
> as a "Dewey Decimal-Like Indexing System" for the Web.
>
> This month's issue of Intelligent Enterprise features an article
> by William J. Lewis which references and explains UDEF.
> I have duly abstracted this in my "XML Articles" collection,
>
>   http://www.oasis-open.org/cover/xmlArticles.html
>
> Or see:
>
> "XML Microstandards"
>  http://www.intelligententerprise.com/000428/supplychain.shtml
>
> - Robin Cover
>
> PS  Actually, I think the *easy* case is when we have
> umpteen names (sic! "tags") for "the same thing", such that a
> solution can amount to a semantic declaration and (numeric/ID)
> "key that allows an unlimited number of aliases."  The
> more common case, and the harder one, is where
> "almost-the-same-thing" (times twelve or thirty) has different
> names AND a slightly different conceptual model; this
> factoring problem is what causes the semantic mismatch,
> and no (easy) magic can be conjured to make the
> same "things" (only 80, 85, 90, 99, 99.2 % "identical")
> behave identically.