RegistryItem/Association: Why PrimaryClass & SubClass

[Len Gallagher <LGallagher@nist.gov>]

RegRep,

Many of you may wonder why two classifications were given special treatment
in the Registry Item entity, i.e. PrimaryClass and SubClassification,
instead of being captured with all of the other potential classifications
in the Classification entity.

The reason, in my mind, is that some classifications are so fundamental to
a Registry specification that they deserve to be treated special.  We can
debate whether this reason is valid. I think the classifications that
determine the "roles" in the Association entity are sufficiently important
to be treated special.  For OASIS registration, I think that the most
important classification is whether a registered item is a "definition" of
an xsgml document, an "instance" of an xsgml document, a "package" of other
registered items, or a "related item".  And secondly, if it is a
"definition" or an "instance", then what kind of xsgml document is it a
definition or instance of.  Then the relationships among these
classifications determine the "uses", "contains", and "related" roles of
the Association entity.

For some other non-OASIS registry, I think the two most important things to
decide are "what gets registered?" and "what are the associations among
registered item?", i.e. the association "roles". Then one can choose the
PrimaryClass and SubClassification definitions to best support those decisions.

One could argue that this is doable even if the PrimaryClass and
SubClassification instances are maintained by the Classification entity. I
just think it's more upfront if they are treated special for each Registry
specification. Others may disagree - in fact, based on the discussion I may
change my mind too, since I find myself wavering every time I think about
it.  However -- if these important classifications get moved to the
Classification entity, then they would need to become required for each
registered item, else the associations we've chosen to maintain wouldn't
make much sense!

-- Len G.




At 08:39 AM 6/23/00 , you wrote:
>Hello All,
>
>At the Paris meeting I took an action item to break the Information Model 
>document into logical pieces that would be useful and manageable for 
>comment and discussion. (Rather than having multiple discussion threads 
>running all over the document.)  The attached file contains that 
>breakdown.  Each item in the list corresponds to a section/paragraph name 
>in the document.  The document is also available at
>http://www.itl.nist.gov/div897/ctg/regrep/breakdown523.html . Here is the 
>first piece:
>
>Registry Item/Associations
>a. Related Enumeration Entities
>	i. Primary Classification (Instance vs Definition)
>	ii. Sub Classification (xsgml types)
>	iii. Association Type
>	iv. Definition Source
>	v. Payment Status
>	vi. Registration Status
>	vii. Stability
>b. Related Entity Semantics
>	i. Registry Item Entity
>	ii. Association Entity
>
>Discussion/comment anyone?
>
>--lisa
>Lisa Carnahan
>National Institute of Standards and Technology
>Information Technology Laboratory
>100 Bureau Drive Stop 8970
>Gaithersburg, Md. 20899
>301-975-3362 voice
>301-948-6213 fax
>lisa.carnahan@nist.gov

**************************************************************
Len Gallagher                             LGallagher@nist.gov
NIST                                      Work: 301-975-3251
Bldg 820  Room 562                        Home: 301-424-1928
Gaithersburg, MD 20899-8970 USA           Fax: 301-948-6213
**************************************************************