On 13/05/2013 20:48, Mitchell, Michael wrote:
… as I understand Bibframe there will no longer be “records. There will be data points and triplets instead. This will be a critical difference and as Deborah says about RDA thinking will be even more true about Bibframe. This frame shift from records to relational data points (I know, I still don’t have the terminology down) is a big reason why I’m so skeptical of anything to do with RDA. I understand that RDA is trying to create rules for more discreet content entry (better data points) but I just think we are spinning our wheels for the most part until Bibframe is closer to development. This is not to take away from the many folks who have been and are working hard on the implementation of RDA but “we’ve” designed a cart before we know if we’re going to hook it to a horse or a jet.
I personally don’t know if it is helpful not to think in terms of “records”. From the public’s point of view, and that of the catalogers and anyone other than a systems person, they will experience a totality of the information associated with a specific information resource, and we will interpret that as a “record”. This is similar to how people refer to a “webpage” when it is now almost never a single file of information, but a unified whole that brings in many, perhaps hundreds, of associated files, and we relate to all of those files as a single page. This is how browsers work. For instance, look at http://www.guardian.co.uk/. See how it loads, how it brings in information and interacts with all kinds of other sites. 99% of this the public does not need to know, e.g. here is only one of their stylesheets that helps the browser display the page correctly: http://static.guim.co.uk/static/f8df97e0df9f9797bda45deb8c5c707dc4e88eaf/common/styles/network-front-grid.css This Guardian page is far from the most complex page that exists. Behind the scenes, it is impossible to predict how everything is structured because there are zillions of ways to decide to display any webpage, plus systems people change those ways constantly for all kinds of reasons, both internal and external.
Opposed to this are the earlier/earliest pages that really were single pages of codes and information, such as Tim Berners-Lee’s original web page: http://info.cern.ch/hypertext/WWW/TheProject.html. The first pages I made were just like his. To see this today is like looking at the stone axes of pre-historic man.
The way catalog records are stored today in a relational database is not single records either. The information we experience as a single bibliographic record is broken into all kinds of tables, each interlinked in a variety of ways. Strangely enough, when it comes to the newest methods that I know of, such as with Lucene indexing, it works only with flat files (in other words, single records more or less as they have been traditionally understood, that Lucene calls “documents”) and there is a completely separate index the computer searches. When someone finds something they want and click on it, they then see the “document”.
One of the reasons I am bringing this up is that for human beings, they will still be experiencing something that displays the information related to a resource. This may be displayed completely or partially. This is exactly what we see today. For instance, if someone searches for the work of Tolstoy’s War and Peace by using the uniform title http://www.worldcat.org/search?q=au%3A%22tolstoy%22+ti%3A%22voina+i+mir%22&qt=results_page you can see it in a multiple display as here, or you can choose to see an individual record. In most catalogs, the individual record can display in an abbreviated form in various ways or in full. This is no different from the way Google and all other search engines work, where people can see the various “hits” (equivalent to multiple display) or click into an individual item (equivalent to record display). It’s also the way FictionFinder worked, which more or less recreated the old book catalog displays except FictionFinder was interactive. (I discussed this earlier http://blog.jweinheimer.net/2009/11/fw-ngc4lib-frbr-wemi-and-identifiers-5.html) These views of individual record/multiple records will probably change very little unless or until somebody comes up with something completely 100% different. Of course that has its own problem: if it is too different then people will probably have troubles understanding it.
The other reason I am bringing this up is because there is absolutely no way that anybody will be able to predict or determine what the structures will be behind the pages/records that people see. Each information system will be built according to its own parameters and those parameters will change.
Therefore, calling them “records” and thinking about them in that way is fine in my opinion, because that is what everyone will continue to experience.