Sunday, September 23, 2012

Re: [RDA-L] Cataloging Matters No. 16

Posting to RDA-L

On 22/09/2012 18:22, Amanda Xu wrote:
For a beautiful Saturday morning in the East Coast, your messages provoked our imagination at work

As far as I am concerned when giving description to the film director of a movie, we just need to markup it using standard vocabularies for data reuse, etc. This can retrospectively make legacy library data compatible to Web of data, Web of software, Web of things, etc.

When adding relators such as Film Director, Author, etc. to a named person in Bibs, we actually create a RDF link that describes the role of the named person for the creation of a film, book, etc. which is critical for discovery in both machine and human processing scenarios of the field.

When this applies to intentional discovery of movies directed by Clint Eastwood over Web, through mapping of the vocabularies from multiple related data sources, the values within selected data sources for the field will be aggregated if data types, constraints of use, provenance matter, etc. are handled in standardized way.

How much details are reasonable for library catalogs? As of June 1, 2012, "Satori [Bing] had mapped over 400 million entities and Knowledge Graph [Google] had reached half a billion, ”a tiny fraction of the potential index of entities that the two search tools could amass." [1]

That's a lot of details for Web search engines. Are we satisfied with such claim as "amass"? Most people would be happy to find a needle in a haystack whether the size of hole of the needle matched to the thread in hand or not.

For professional catalogers, I would hope through RDA/FRBR training, the profession would do a better job tailoring what we do to meet the need of user communities that we are serving by harmonizing vocabularies and tools provided by search engines to such an extent that it actually enriches library catalogs as inventory, entry points, etc. at field level.

Many talents on this list can echo their experience in metadata applications for E-Resource management and why relators are important for discovery and whatever SW stuff is required underneath it.

Should we do less or more? As far as I am concerned, metadata quality has always been affected by changes, let alone other factors such as data quality and consistency. If "being less" yet "being the truth", and others can rely upon what we produce for reuse, then "being less" is good enough. However, our systems have to be flexible enough to compose and satisfy the intentional discovery need for our users on demand. Which way to go is obvious to us.

Thank you for your comments. The problem, as I see it, is something different however. I agree that simply adding the code to a person for the role information (actor, editor, etc.) is not that much extra work for the cataloger (most of the time!), although in the aggregate, it will definitely decrease productivity. This is not the problem I am pointing out. The simple fact is that if a search for "film directors" is implemented, the result for the user of the catalog will be that access will go down tremendously, and this is because of lack of consistency.

There are a few points of reality that need to always be kept in mind. They may not make a cataloger very happy, but they are facts.

When someone comes to a library, they do not come for the library's tools; they do not come for the librarians. They come to the library for its resources: its books and journal articles and maps and films and everything else. The public will use the library's tools (and the librarians themselves!) only as intermediaries to get at the resources they want. They want to spend the least amount of their time in the catalog reading the records and/or fighting with it because they are interested in the information in the books and journals and films, etc. This includes me when I am using the catalog as a researcher.

In this sense, when people approach the library's tools, they are similar to people who are building something, e.g. a room onto their house. They will need several tools, including, let's say, a power saw. People are not interested in the power saw in itself--they just want to use the power saw to help them build the room onto their house. Most don't want to spend time learning about the power saw, how best to work with it and so on. It is a tool that is used for their own purposes.

If that tool changes (as has happened many times with many tools when you replace your old tools), as people use the new saw in the old ways, they will often discover that it doesn't saw as well for some reason. This is because it has been redesigned and they have to relearn how to use the tool. Some do not succeed and the new saw remains unused. Still, the changes may have advantages as well as disadvantages.

If RDA is to be aimed at the public, we must first of all consider the impacts on them. Unfortunately, there has been very little substantive research on how people will respond to the RDA changes, so it is all a mystery at this point. In my latest podcast, I demonstrated how RDA will lead to decreased access. That is a fact because of the fundamental principles of what makes a catalog a catalog. What will happen?

The public will discover that the catalog doesn't retrieve as much as it used to. It will become even stranger than it is now and they will quickly see the "errors". "I looked for D.W. Griffith as film director and got nothing. Then I looked for him *not* as a film director and got lots of stuff. Those catalogers are worthless! You can't trust anything in that catalog..." This is a normal human reaction and can be predicted very easily by anybody.

There is a wonderful book by Donald Norman, "The design of everyday things" that opened my eyes to the world of "user-centered design". For instance, in his book he has an amazing discussion about doors(!): he admits that he has a lot of trouble with doors: pushing when he should pull, or opening them the wrong way, and he talks about how and why some doors work well and others don't. He also mentions that when he stayed in a room at a university or hotel, the sink for the water came with a user's manual! You had to read the manual before you could get any water! That's crazy! He also mentions that if you are lucky enough to buy a new tool and discover that it fits in your hand perfectly and you can use it immediately, you mostly don't think about it, but it is evidence of a lot of hard work and absolutely brilliant design. It is a tremendous book and I suggest it to everyone, cataloger, librarian, or not.

It is my fervent belief that what we need to do is build something that follows this "user-centered design." It means exactly what it says, and I have seen little evidence of this with RDA. One example is no concern for the lack of consistency in the catalog. What is the user to do? Maybe in an ideal world, adding "film director" would be better for our public and maybe not (no research has been done) but in the world we actually live in, people will retrieve a fraction of the materials really available to them, and if they are not to retrieve very little, the final product of adding the relator codes to our new records will be similar to expecting people to read a user's manual to get water out of the tap. If searchers are to keep from missing 99% of everything that is available when they search for film directors, they are going to have to know a lot, and that means reading and understanding the user's manual.

Everybody knows the public won't do that and will find our catalogs less useful than ever. And we must remember, in our case, the public will have to read the user's manual *not* to get the the actual resources they want (at least if you read the user's manual for the sink, you would actually get the water), but in the case of the catalog, people will need to learn how to use the tool just *to find* the books and materials that they will then have to obtain. RDA does not talk about the consequences of decreased access, but librarians and the public will have to deal with it from the first day such a search option is implemented. What is everyone to do?

There are too many alternatives out there for people today to find information. They will happily use tools that are easier. I have mentioned plenty of places where the catalog could be improved without changing a single cataloging rule. But I have stated these enough times already.

No comments:

Post a Comment