I have a few brief comments on the May 14 authorities proposal update:

1) the naming of "hasAuthority" and then "referenceAuthority" seems inconsistent because one uses "has" and the other does not, yet they are conceptually similar. Most of BIBFRAME does not use the "hasX" name form. [1] "Authority" alone seems understandable.

2) there is no mention of authorities in multiple languages. It seems that this needs to follow the SKOS convention of a single "prefLabel" *per language*. This could translate either to more than one authorized access point, or more than one hasAuthority.

3) for works as subjects, the assumption is that there will always be a URI for the work. RDA allows a work to be described either with an identifier or a citation. It does seem that there will be cases in which there is no URI for the work as subject. Perhaps an example could show how to do this?

4) I note that in your examples, at no point are the BIBFRAME authorities given "typeOf" BIBFRAME authority. In essence, they have become nodes that can be of type Person, Work, Place, etc. Whether or not they are "authoritative" will depend on whether an authority is referenced in the node. Is this your intention?

5) Could there be both an authorizedAccessPoint and a label in the same node? e.g.

    <bf:authorizedAccessPoint>  Cutright, Paul Russell, 1897-</bf:authorizedAccessPoint>
    <bf:label>Paul Russell Cutright</bf:label>

[1] http://www.w3.org/wiki/RoleNoun
Karen Coyle
[log in to unmask] http://kcoyle.net
m: 1-510-435-8234
skype: kcoylenet