Quoting John Attig <jx...@psu.edu>:
There are many things that can act in more than one way in a FRBR
model. One of these is Place, which is a group 3 entity, but is
also an attribute of several other entities: Place of publication,
Place associated with a person or corporate body, etc. The FRBR
working group made a conscious decision NOT to model all of these as
relationships to the Place entity.
John,
It may seem logical that "Place of publication" and the group 3 entity
"Place" are related, but that is not the case in the FRBR E-R model.
The fact that the word or even the concept of Place is used elsewhere
in FRBR does *not* create a relationship in an E-R model, and no
formal relationship is defined between those attributes and the Place
entity. In fact, the FRBR entity Place can only be used as a subject
as it is currently defined. Things cannot act in more than one way
unless they are formally defined in more than one way. E-R modeling
does not allow for other interpretations.
If that is not the intention of the creators of FRBR, they should
allow others with greater knowledge of data modeling to re-design the
FRBR model such that it meets the needs of the community. Place could
be defined as a general entity that can be used in various contexts.
However, in my experience any attempts to create a workable model that
varies from the letter of FRBR has met great resistance. If the model
is broken it needs to be fixed.
kc
The case with Form/Genre is similar. There are a number of
attributes that relate to the form of the entity. However, a
form/genre entity currently lies outside the scope of the Functional
Requirements model. It was omitted from both FRBR and FRAD and was
not included in the scope of the proposed FRSAR model.
There is, however, an excellent analysis by Tom Delsey, which I
regularly commend to people's attention. In a presentation in 2005,
Tom considered "Modelling subject access : refining and extending
the FRBR and FRAR conceptual models". Slide 18 is a redrawing of the
FRBR entity-relationship diagram to show Tom's proposed new entities
and relationships. Prominent among them is "Form/Genre" which he
defines in a "is example of" relationship with the Work entity
[although I would argue that the relationship could be to any of the
group 1 entities]. There are other new entities and relationships,
including a "coverage" relationships between Work and two entities
that Tom calls "Time" and "State". This is a fascinating exercise
in data modelling and reminds us that work on the Functional
Requirements models is not a finished exercise, but that significant
extensions still need to be undertaken.
Again, I recommend this presentation to anyone who is thinking about
going beyond the current state of the models.
John
--
Karen Coyle
kco...@kcoyle.net http://kcoyle.net
ph: 1-510-540-7596
m: 1-510-435-8234
skype: kcoylenet
