OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

regrep-semantic message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: notes on revision of April 19 Use Cases



Notes on V2 use cases:

I have made fairly minor revisions to the use cases.  Most of the
changes were editorial, not significantly changing the content.  I
turned on change tracking partly into editing use case 1 (sorry).
IMHO still more work is needed on this document before releasing it
to outside review.  One additional change I would like to see would be
the inclusion of an overview explaining how each use case fits into
the overall picture.  For instance the first two represent uses of
Semantic Content that could be implemented with current RIM taxonomy
support, but would be better served with richer ontology support.
Also none of the use case diagrams originally created have been
imported into the document (or perhaps that was just a function of the
RTF version I was working from).

Below are some notes (mostly) created while editing.  

Use Case 1:

Tried to expand most acronyms that wouldn't be obvious to external
readers.  "STEP" isn't an organization, so I replaced it with the ISO
subcommittee that produced the standard.  I guessed at other acronyms.
David should verify that I got these right.  Also not sure what the
proper expansion is for EPR.  I also took some liberties revising the
text to better match its role as an SCM use case.

Use Case 2:

Changes similar to 1.

Use Case 3:

The addition of the assumed existence of configuration management
policies was the only significant change to this use case.

Use Case 4:

Added further explanation of the benefit of supporting this use case with RIM.

Use Case 5: 

I am not sure that I understand use case 5.  A class in an ontology is
not an appropriate analog to an Association.  An association
(assuming UML-like semantics for association) would map into a
property and its inverse, with domains and ranges of those properties
restricted to the object classes associated.  This is obviously not a
very simple mapping and requires something more than an analog to
ClassificationNode.  This is even worse if the association can specify
multiplicities (I have worked up a whole set of rules for this sort of
mapping from UML to OWL).

This leads me to believe that the having a separate use case for this
IS a good idea.  I have changed the use case accordingly to match my
current understanding.

Use Case 6:

I am not sure that I understand this one either.  Is it going to be
replaced by use cases provided by Jeff anyway?  It seems to me that
there are a number of kinds of content that a semantic query could
retrieve: 1) semantic content a) classes or b) instances, 2) schema
elements corresponding to semantic a) classes or b) instances, or 3)
data elements representing instances of a semantic class.  I think
that SCM RIM enhancements should support all of these.

Use Case 7:

This use case is about providing a stronger means of user
extensibility of the RIM metamodel, correct?  Is it a fair assumption
that ontology definitions would exist in this use case for all
metaclasses in the RIM?  Didn't change this one.

Use Case 8:

This is now cleaned up a bit.  It might be easier to understand if a
different query syntax were used.  Suggestions are welcome.  In any
case, it should be updated when 6 is reworked (I think).

Use Case 9:

I think that the example is a different (and even more interesting - 
IMHO) case than the overall Use Case seems to be addressing.  The
relationship between Winter Package and Heated Seats seems to be a
hasComponents relationship rather than equivalence.  This is one
example of why we need a facility as now described in Use Case 5.
s


-Evan


[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]