Well, I finally got around to reading the PCC report on Expression Task Group. Read it on paper; a bit hard to get through it online.
Considering what caused a lot of discussion to arise, a few comments first.
Looking at the list of librarians that made up the report, I would suggest the group consult other organizations whose publications they publish or study also merit special consideration. A few come to mind: Society for Classical Studies, and the North American Patristic Society. These societies deal with works that are not "published" by current means, whose characteristics, construction are very different from modern publishing methods.
I was reading it and I kept getting the feeling that we weren't going anywhere. And then I came to page 7, where I realized that the task group was leaving in the same mud that some of us have been talking about. Basically, pick one method, RDA or LC, and be consistent. Cataloger's judgment.
Just some specific comments. On page 15, we have the statement since the NARs and bib records are routinely distributed to OCLC and Skyriver, therefore LC is considered the database of record for deciding whether or not it is necessary to create an NAR for the work or expression. That seems to contradict what it says in the first part of #3 and in #2. It would seem that if we are cataloging items that are not yet in either of those databases, either bibliographically or in the NARs, we would have an obligation to create an NAR.
On page 16, #6, example one: it is not clear why creating an NAR for the Spanish expression would be redundant. To what? The Spanish of the title in the 245? Who would know that without the 240, which I assume would be in the NAF.
Page 17, #7, nos. 2-3. Why would you not create an NAR for the work since the expression is of that work? If we only create the expression, how can we be sure that when the work NAR is created that it will have the same form, minus the $l, that the expression has? Note that there is comment that different criteria apply to musical and legal resources. Same goes for classical and some medieval works.
p.18, continuation of #8. First of all, one has to recognize that while the personal names may be controlled here, the titles are not. And what is not discussed here and what got the whole previous discussion are works that do not have a collective title; the title is made up of a listing of the titles that are in the work; one sees this a good deal in publications of classical works.
And perhaps there should be an FAQ for such works.
Anyway, this report should refined and acted upon so that the instructions will be more clearly formulated for catalogers who wish to create consistent, authoritative entries to the resources they are cataloging.