Karen Coyle wrote:
>> I'm not sure how the roles are treated -- whether they can be 
>> interpreted
>> as elements in their own right, like publisher, or not. In other
>> words, could you make a logical list of relationships that has:
>>
>> creator
>> illustrator
>> publisher
>> translator
>>
>> even though some of these may be seen as more "essential" in the
>> bibliographic description, and others as somewhat secondary?
> John Attig wrote:
> I can't give you a technical answer, but what you have done makes 
> sense to me.  A logical list of valid relationships includes both the 
> relationship elements and the subordinate roles defined in the 
> Appendix.  The listing and use of such a logical list (it seems to me) 
> is an implementation decision. We chose not to formally define each 
> role as an element sub-type of the relationship element (mainly to 
> avoid another volume of RDA text), but our understanding has always 
> been that the roles are in fact element sub-types.  In the registry, 
> could you designate these as sub-types and indicate which relationship 
> elements they are subordinate to?  That seems an ideal solution to me.
>
Yes, if you look at the MARC Relators that have been defined as 
sub-properties of DC element Contributor (look here: 
http://www.loc.gov/loc.terms/relators/dc-relators.xml), you can see that 
the relationships are built into the formal representation.  This is the 
idea with the RDA Roles, but as the FRBR registrations have yet been 
done (Gordon is doing that, last I heard), we have yet to do that part.  
I'm not sure that I quite understand the idea of roles being 
sub-properties of relationships, but we'll jump off that bridge when we 
come to it (much will depend on the definitions of the relationships, 
and how they coordinate with the definitions of the roles).
> One further distinction may be important: that between the 
> relationship (which RDA treats as the related entity -- related 
> person, related corporate body, related work, etc.) and the 
> relationship designator.  Creator and publisher above are related 
> entities, while illustrator and translator are designations associated 
> with a related entity.  In the case of creator and publisher, no 
> designation is necessary because the nature of the relationships is 
> inherent in the element. In practice, this distinction is probably 
> based on the "librarian's" view of the anticipated data structure that 
> you referred to in your previous message.
>
I'm not sure I fully "get" these distinctions either, but hopefully 
we'll manage to figure that out as well. 

Diane