It could be reasonably asked why this is in text when it could have been on the podcast. There are reasons for such. Unfortunately I cannot let loose with spoilers at this time.
Norman Oder reports at LJ online about Karen Calhoun speaking at ALA. I wish I could have been there. I do not relish the thought of asking on AUTOCAT how acrimonious the session was or was not.
The OCLC data policy has created quite a bit of consternation. This is understandable. When I was last working in libraries, my specialty is as a cataloger. I have worked at an OCLC member and at a non-member that relied heavily on Z39.50. There are advantages and disadvantages to both. I've cataloged under both and have real-world experience in the past three years.
In this age of community, WorldCat is not the appropriate mechanism. There is plenty of utter crap in WorldCat that libraries freely share. The structure of cooperative cataloging now more closely resembles a command-and-control economy rather than a community. Outside AUTOCAT, catalogers cannot interact. WorldCat's structure already makes it quite difficult to remove errors. This new data policy has the practical effect of making mistakes immortal.
Z39.50 by itself does not allow interaction either. Used in conjunction with a listserv, though, collaboration can be possible. Z39.50 also allows for quality control by voting between differing record versions to find the one that fits best. Working this way would force the fostering of community as trust in records would be built on the reputations of those creating them. There would be no red tape to hide behind if you were not able to catalog materials up to par.
I am torn. In some respects, OCLC really is not fit for purpose as presently constituted. Years of mission creep have resulted in an agglomeration of functions that are not completely connected. Stripping away functions to where OCLC's only purpose is resource sharing would be a good step. The research office as well as the offices of Roy Tennant, Karen Calhoun, and Lorcan Dempsey are nice but are more appropriate in an academic setting or think-tank than where they are now. The contract services portions could just as easily be sold off to an established commercial vendor or two. OCLC is already going madly off in many directions while focus on core functionality is lacking.
My favored solution that few would likely agree with would involve a dismembering of OCLC and an expansion of the Program for Cooperative Cataloging. The PCC already offers some means for communication and interaction. As an umbrella to more local community networks it would allow for potentially simpler structure of a community. Similar organizations already exist with such structures like the American Radio Relay League and the International Amateur Radio Union. This would not be an earth-shattering kaboom but something that has precedents elsewhere.
If memory serves, there is nothing internal to WorldCat in MARC21 but in Dublin Core instead. For the purposes of inter-library resource sharing, minimal-level records would conceivably be sufficient. Full records would only be necessary at the local level. The minimal-level records in WorldCat.org would make such a topological shift fairly simple to implement with minimal records linking to more full local records.
The data policy revisions are not a case of cranky catalogers grumbling at other folks in the cataloging realm. The policy instead relates to the topology of the knowledge ecology all librarians care for. OCLC's changes have implications not only for catalogers but also front-line reference librarians. If OCLC owns the data and provides its own discovery tools like WorldCat.org, what manager in this economy could justify the outlay of a salary for a degreed professional that would seem to be duplicating a "free" web tool?
I have thrown some ideas out in this post of ways things could adapt to our changing economy. I do not doubt somebody might feel such to be wrong or ill-informed. Hopefully some discussion can result from this.