On 12/5/2013 4:30 PM, Dan Matei wrote:
<snip>Friend Dan, (and friends can always disagree!)
... who is thinking that FRBR is the "ultimate solution". I did not meet anyone from that sect !
My impression is that you could save some effort on your crusade against FRBR :-) It is not the main enemy!
Please, do me a favor: search "Hamlet" in Google. What do you see right on top ? And what is that ? The wikipedians are doing FRBR without even knowing it ? As M. Jourdain with his prose.
BTW. Please search also "M. Jourdain" :-)
I do believe that FRBR is "the main enemy" (to use your term). Why? Because everything, including RDA and the new formats, etc. all state explicitly that this is what they are aiming for, even though the model has never been proven to be what people want. Why should we assume that people want it? On the contrary, is there any evidence that people do not want FRBR? Yes, and it is highly significant.
The structure of the card catalog allowed people to do the FRBR user tasks (where--for those who understood--people really and truly could find/identify/select/obtain works/expressions/manifestation/items by their authors/titles/subjects (or at least they could if the catalogers had done their jobs correctly). All of that was codified in Cutter's Rules. As a result, people really and truly could do all of that in the card catalogs and that can be demonstrated. I have done so myself. And yet, the moment that keyword was implemented in online catalogs, all of that vanished. What was the result? I never heard anyone complain--except a few catalogers (such as myself) and a few reference librarians once in awhile. Everyone else was only happy. Many times I would point out to people, "But you are missing this, and this, and this" and the reply was, "So what? This is such a great improvement!" My efforts were not appreciated. And that was it. As time went on, the few who complained stopped complaining (including myself) and moved on to other structures that people obviously preferred.
There has been a major change since that time. The fact is right now faceted catalogs allow the FRBR user tasks again, and they do it far more than the Hamlet example in Google shows. For instance, we can see it with the search in Worldcat for au:"Shakespeare William" and ti:Hamlet http://www.worldcat.org/search?q=au%3A%22shakespeare+william%22+ti%3Ahamlet&qt=results_page which shows the entire work of Shakespeare's Hamlet (once again, this statement is true only if catalogers are doing their jobs correctly), and once you have done this search, by using the facets in the left menu, we can navigate by language, by dates of publication, by other authors etc. Of course, we could implement more facets like these if we wanted.
I admit the current user interface isn't very clear for the non-initiated, but that can be improved too. (I've given a paper on that, by the way) So, if the FRBR user tasks really and truly are so important and so necessary, the unavoidable fact is: they have already been achieved. We can do it all right now. And all it took was some brilliant programming. No records need be harmed in the making of FRBR. :-)
But the overwhelming silence that has greeted this development (which I think is one of the greatest advances in the history of the catalog) makes me more skeptical than ever that FRBR is what people want. So then the argument morphs into: now we need new relationships to be entered into the records, so that searchers can see when something is an adaptation, as opposed to a parody or a summary, or whatever. Of course, to get those relationships to work as access, those relationships need to be added to the "legacy data".
What does that actually mean? Well, I checked in the latest MARC usage in Worldcat (I sent this discussion to Autocat but it's also in my blog) and discovered that for 700$t (which I assume will need to be reviewed for possible relationship information) there are 6,617,885 instances. 240$a has 9,137,082 and I would assume at least some of these would need some level of review. The number of 730$a is 4,877,321, and at least some of these I assume would need a review also. There are also 710$t but those are less than a million. Still, all of those numbers add up to quite a lot of records and even if some of it could be automated, such that perhaps as much as 30% could be done automatically, that still leaves rather a lot. I haven't had the courage to look at what the 100/700$e would entail.
So, what is supposed to happen with these millions of records? Are we supposed to hire thousands of catalogers to do them? Who will pay for them? Do we think we can crowdsource it? Adding those relationships is not simple, as questions on this list and other lists have shown. Are we praying that no one will notice? Nobody has said anything about this. There is still a lingering suspicion in my mind that the normal IT solution will be implemented: just throw those records in a .zip file and put them where people can download them if they want.(!)
These developments are driven by FRBR, even though it has never been proven to be what people want (and as the paper by Amanda Cossham shows, I am far from alone saying this). FRBR has certainly not been taken up by other institutions. One of the main purposes of RDA was to implement FRBR.
Shouldn't the pro-FRBR people have to demonstrate how important all of this is to the users? Or has the argument morphed again into "linked data"--although this comes up against yet another uncomfortable fact: linked data does not need FRBR either.
So yes, FRBR is certainly a major problem. It is a discussion that should have taken place quite a few years ago. But, better late than never....