Thank you Owen! these is all the questions I've had but you've gone 
through them more thoroughly.

 > Interesting though all this discussion is, I'm not clear how it moves us
 > forward. The discussion developed out of Pete's question about using
 > 'concept' as opposed to other types for the vocabularies. I'm trying to
 > get my head round a lot of this for the first time, so apologies if this
 > is just wrong, but as far as I can see SKOS essentially works with
 > 'concepts' - there aren't any other choices, so as long as we are using
 > SKOS for vocabularies, we are using concepts - have I missed something
 > here? SKOS seems appropriate for this kind of limited vocabulary
 > listing. However, it would seem inappropriate to try to use it for
 > persons or corporate bodies (and possibly other complex knowledge). As
 > far as I can see, noone has suggested this so far?

Owen, we are only using SKOS for the value vocabularies, not the RDA 
"elements". The question Pete brought up was whether we consider RDA 
term lists like "base material" to be concepts or "things". At this 
point, I think that separating the many and varied RDA value lists into 
those that are concepts and those that are things would be very 
difficult (I'm sure we'd find one that is a combination of the two!), so 
my preference is to code them all in SKOS for now (we have a deadline) 
so that RDA work can move forward.

kc

Stephens, Owen wrote:
> OK - I seem to have written a rather long and rambling email here, so
> for those who want the quick version, the summary is
> 
> *	FRBR and FRAD 'person' entities are fundamentally different
> *	RDA maps only to the FRAD 'person' entity, so this is the only
> one we need to examine (although I have some real problems with the FRAD
> definitions)
> *	FRAD 'person' entity represents a 'persona' OR a 'person', so we
> can only ever assume a specific entity is a 'persona' (in some cases it
> may be a 'person')
> *	FRAD defines person to person entity relationships - so we can
> show a person entity that is a pseudonym for another person entity
> *	FRAD is clear that you can have a 'person' entity that
> represents two authors writing under a single name. However, it is not
> clear how this would be represented in terms of relationships between
> entities (pseudonymous, collaboration or both?)
> *	Under the current FRAD model the only way (I can see) of
> addressing Karen's concern would be for every person entity to have at
> least one 'pseudonymous' relationship to another person entity with one
> representing the person and one representing the persona of that person
> when authoring creative works - but this would seem like doubling our
> work to accommodate edge cases?
> *	In response to Pete's statement "(I think) all instances of FOAF
> Person are instances of FRBR Person, but not the other way round." - for
> RDA/DC work we need to consider rather the relationship to the FRAD
> Person, but it holds true - all instances of FOAF Person and instances
> of FRAD person, but not the other way round.
> *	SKOS uses concepts to represent stuff - it's how it works
> (please correct me if I'm wrong)
> *	SKOS is OK for controlled vocabularies of the type described
> here. I'm not sure it would work for more complex data such as People,
> Events, etc. - but also I'm not sure there is any suggestion that it
> should be used in this context?
> *	I wish there was a Dummy's Guide to RDA/FRBR/FRAD/DCMI as it
> leaves me confused.
> 
> For those who want to see my working, keep reading
> ------------------------------------------------------------------------
> --------------------------
> 
>>> But we seem to 
>>> have established that the FRAD person and the 
>>> FRBR person have different qualities.
>> I don't think that is what is intended.  My 
>> understanding is that the group 2 entities in 
>> FRAD are supposed to be identical with the same 
>> entities in FRBR.  I think we chose to map the 
>> FRAD entities because FRAD contains the most 
>> up-to-date list of attributes for those entities.
> 
> It may not have been what was intended, but as far as I can see this is
> exactly what has happened.
> 
> I think FRBR is pretty weak in terms of dealing with the person entity.
> It clearly states in 3.2.5:
> "Defining the entity person enables us to name and identify the
> individual in a consistent manner, independently of how the individual's
> name appears on or in any particular expression or manifestation of a
> work."
> 
> Use of 'individual' and 'person' (who have birth and death dates as
> attributes) clearly suggests we are talking about biological persons. 
> 
> At the same time it says in 4.6.1:
> 
> "In some cases (e.g., in the case of a person who writes under more than
> one pseudonym, or a person who writes both in an official capacity and
> as an individual) the bibliographic agency may establish more than one
> uniform heading for the person."
> 
> This is is a mess I think - what does it mean to 'establish more than
> one uniform heading' - does this suggest a separate entity or simply a
> attribute of the existing entity. Since attributes don't have a
> qualifier of 'uniform heading' I'm left completely unclear as to what is
> going on here.
> 
> FRBR says nothing about relationships between 'person' entities which
> would mean that having separate entities for pseudonyms would leave us
> unable to show that one 'person' entity was a pseudonym for another.
> 
> FRBR defines a 'corporate body' in 3.2.6
> 
> "The entity defined as corporate body encompasses organizations and
> groups of individuals and/or organizations that are identified by a
> particular name"
> 
> Which would deal with the 'two authors writing under a single name' - it
> seems clear that FRBR sees this as a corporate body (it is very clear
> that a 'person' is an 'individual') (which I think is both sensible and
> straightforward)
> 
> FRAD on the otherhand immediately confuses the issue by stating
> explicitly that two authors writing under a single name are a 'person'.
> Since it then goes on to say that a 'Corporate Body' "includes
> organizations and groups of individuals ... identified by a particular
> name", it seems to me that it has immediately muddied the waters,
> leaving confused definitions which overlap.
> 
> What FRAD is very clear about is that the FRAD 'Person' entity can
> represent both "real individuals" and "personas established or adopted
> by an individual through the use of more than one name (e.g., the
> individual's real name and/or one or more pseudonyms)."
> 
> Since RDA maps to the FRAD entity, and not the FRBR entity, I think the
> answer to Karen's question:
> 
>> 2. You now have "names," "personas" and "persons". And personas =
> persons. Which of these is = FRBR:person?
> 
> is, in terms of RDA, none - RDA only maps to the FRAD entity which is
> called 'person' but represents 'personas'. 
> 
> I can see why RDA has mapped to the FRAD person entity - FRAD contains
> much more useful detail, and specifies relationships between person
> entities. This shows us that we can have a 'pseudonymous' relationship,
> so we now have a way of relating a persona to a person. However, it is
> not clear to me how a persona representing multiple persons would be
> represented in terms of relationships - is this done via pseudonymous or
> collaborative relationships, or both?
> 
> One way of addressing Karen's concern within FRAD would be to argue that
> you could always represent a 'person as author' separately to the
> 'person' (both as person entities) and create a pseudonymous
> relationship between them. Actually at the weekend I noticed the new
> James Bond book out on the shelves. It was authored by 'Sebastian Faulks
> writing as Ian Fleming' - which suggests that at least in this instance
> there is a concept of a 'persona' of Ian Fleming who writes James Bond,
> and two people have now inhabited this persona - Ian Fleming and
> Sebastian Faulks. I'd suggest the best way of representing this in FRAD
> would be to create a person entity the following:
> 
> Sebastian Faulks
> Ian Fleming
> Ian Fleming ('as persona who writes James Bond')
> 
> And create pseudonymous relationships between the first two and the
> third. We would also need (of course) a "Sebastian Faulks ('as author')"
> entity with a pseudonymous relationship to Sebastian Faulks.
> 
> In response to Pete's statement "(I think) all instances of FOAF Person
> are instances of FRBR Person, but not the other way round." - for RDA/DC
> work we need to consider rather the relationship to the FRAD Person.
> FOAF says
> 
> "The foaf:Person class represents people. Something is a foaf:Person if
> it is a person. We don't nitpic about whether they're alive, dead, real,
> or imaginary."
> 
> This doesn't include 'groups' or 'organisations', or 'software'. All
> this suggests that all instances of FOAF person are instances of FRAD
> person but not the other way round (a FRAD person could be a
> foaf:person, foaf:group or a foaf:agent; also a FRAD corporate body
> could be a foaf:group, foaf:organisation or a foaf:agent)
> 
> Interesting though all this discussion is, I'm not clear how it moves us
> forward. The discussion developed out of Pete's question about using
> 'concept' as opposed to other types for the vocabularies. I'm trying to
> get my head round a lot of this for the first time, so apologies if this
> is just wrong, but as far as I can see SKOS essentially works with
> 'concepts' - there aren't any other choices, so as long as we are using
> SKOS for vocabularies, we are using concepts - have I missed something
> here? SKOS seems appropriate for this kind of limited vocabulary
> listing. However, it would seem inappropriate to try to use it for
> persons or corporate bodies (and possibly other complex knowledge). As
> far as I can see, noone has suggested this so far?
> 
> I still find myself confused by the relationships between various parts
> of this puzzle. I wish there was somekind of 'for dummies' guide!
> 
> Enough for today
> 
> Owen
> 
> 
>> FRBR 4.6.1 isn't really talking about 
>> bibliographic identities.  This is more about 
>> name variations: this includes variations of the 
>> same name (initials vs. full names, etc.) and 
>> different names (the most common case is a person 
>> who changes his or her name, e.g. a married name) 
>> -- not when a single biological person adopts 
>> distinct identities as the creator or contributor 
>> to a bibliographic resource.  RDA is consistent 
>> with 4.6.1, but does not consider pseudonyms to 
>> fall under these guidelines; rather RDA 
>> interprets the definition of person to cover our 
>> concept of separate bibliographic identities -- 
>> thus treating them as distinct persons.
>>
>>> I note that the RDA mapping to FRBR does not 
>>> include FRBR person; and that the RDA mapping to 
>>> FRAD includes the FRAD person. But we seem to 
>>> have established that the FRAD person and the 
>>> FRBR person have different qualities.
>> I don't think that is what is intended.  My 
>> understanding is that the group 2 entities in 
>> FRAD are supposed to be identical with the same 
>> entities in FRBR.  I think we chose to map the 
>> FRAD entities because FRAD contains the most 
>> up-to-date list of attributes for those entities.
>>
>>> The FRAD (FRANAR) definition of a person is:
>>>
>>> "Person
>>> An individual or a persona established or 
>>> adopted by an individual or group. [FRBR, modified]
>>>
>>> Includes real individuals. Includes personas 
>>> established or adopted by an individual through 
>>> the use of more than one name (e.g., the 
>>> individual's real name and/or one or more 
>>> pseudonyms). Includes personas established or 
>>> adopted jointly by two or more individuals 
>>> (e.g., Ellery Queen - joint pseudonym of 
>>> Frederic Dannay and Manfred B. Lee). Includes 
>>> personas established or adopted by a group (e.g., Betty Crocker)."
>>>
>>> This is a different definition from the FRBR one 
>>> (and it even says so here). So if I were to 
>>> define Person as an entity for the purposes of 
>>> RDA, it looks like I would need to use FRAD as 
>>> my basis, not FRBR. *OR* is FRBR being modified 
>>> to use the FRAD definition? Are there other 
>>> entities for which we should use the FRAD 
>>> definition and not the FRBR definition?
>> I would interpret this as an extension of FRBR, 
>> rather than a substantive difference; as I tried 
>> to indicate about FRBR 4.6.1, I don't think that 
>> the FRAD text contradicts FRBR, it simply adds 
>> something not explicit in FRBR.  I would expect 
>> that someday the two models will be reconciled 
>> and I would expect that the FRAD text would be 
>> added to FRBR -- or both models modified in a 
>> consistent way.  This is an inevitable problem 
>> when two different groups working separately and 
>> in succession look at what are supposed to be the 
>> same things.  And there is a third group working 
>> on subject authorities, which (from indirect 
>> reports) is taking an approach different from either FRBR or FRAD.
>>
>> So, in terms of the definition as well as the 
>> attributes, I think that RDA is based on FRAD rather than FRBR.
>>
>> Beyond that, I think that RDA does not adopt all 
>> aspects of FRAD.  As I have been trying to argue, 
>> RDA is not explicitly about authority control of 
>> access points for names of entities.  RDA does 
>> not include the Name entity (we treat name as an 
>> attribute rather than an entity, as does FRBR [I 
>> think]) nor the Rules or Agency 
>> entities.  Personally I believe that authority 
>> records have a place in a relational structure 
>> designed to support RDA, but the relationships 
>> are different and more complicated than those in the FRAD model.
>>
>> And, of course, we are all making assumptions 
>> about what FRAD will actually say when it is 
>> published.  The length of time it is taking to 
>> finalize the document indicates that there are 
>> still issues under discussion and changes being considered.
>>
>>          John Attig
>>          Authority Control Librarian
>>          Penn State University
>>
> 
> 

-- 
-----------------------------------
Karen Coyle / Digital Library Consultant
[log in to unmask] http://www.kcoyle.net
ph.: 510-540-7596   skype: kcoylenet
fx.: 510-848-3913
mo.: 510-435-8234
------------------------------------