Re: [RDA-L] Qualifying access points

Posting to RDA-L

On 04/11/2013 22.49, Brenndorfer, Thomas wrote:

<snip>
If catalogs can’t take people to authority records (and some can), Wikipedia doesn’t seem to mind. It’s just a question of programmers matching the data to the users. Here are some examples of what’s possible when one sees the forest of possibilities:

Authority data links from:
http://en.wikipedia.org/wiki/Elvis_presley

German DNB authority record for Elvis Presley (with all the RDA equivalent elements wonderfully accessible for any catalog user):
http://d-nb.info/gnd/118596357/about/html

WorldCat Identities (lots of attribute and relationship elements here, nicely meshed together without much thought to restrictions based on what can fit into a catalog based on 5X3 cards):
http://www.worldcat.org/identities/lccn-n78-79487

LC authority data in id.loc.gov:
http://id.loc.gov/authorities/names/n78079487.html
And more RDA goodness here and ready to be integrated when systems are ready: http://lccn.loc.gov/n78079487
</snip>

Looks like something only a cataloger could love. ūüôā

Of course, Wikipedia itself is not linked data, but dbpedia is derived from it and the actual linked data is at:
http://dbpedia.org/page/Elvis_Presley

Whenever I find myself examining those sites you mention (and there are tons more linked data nodes besides these), I am surprised by the amount of duplicated effort. Some appear to be just different views of the same information, e.g. http://id.loc.gov/authorities/names/n78079487.html and http://lccn.loc.gov/n78079487, but others seem to be made separately. For instance, do the library authority files simply repeat bits that are in the huge dbpedia page? I don’t know but it wouldn’t surprise me.

The most useful of the sites you mention (besides Wikipedia) seems to me to be Worldcat Identities because it appears to create something never seen before. And yet, when I have shown this to people (many times, actually), they are impressed but have never seen how it could be useful for their own needs. I have thought that perhaps the word cloud at the bottom could be useful, but as implemented now, it is not. Perhaps if the links there included Elvis’ heading so that when you click on one, you could search Presley, Elvis 1935-1977 with e.g. “popular culture” it may be more useful. Or not.

Plus there are some curious glitches. For instance, I was surprised that Worldcat Identities claims that the most widely held work by Elvis Presley is Lilo and Stitch, something I had never heard of. It turns out that it is a cartoon that came out long after his death. I have never seen it, but Elvis apparently does not make an appearance, although some of his songs are in it. (http://en.wikipedia.org/wiki/Lilo_&_Stitch)

The purpose of authority files has been to help people search related catalogs more effectively by providing a series of cross-references. That was all they were there for. Authority files are valid only for their own catalogs and do not work in catalogs with other rules. VIAF may change that. Being able to use the cross-references to help you search correctly is a critical function that was lost with the introduction of keyword searching.

To expect more of authority files is to take them beyond their function and we should not just assume that bringing it all together will be useful. To determine whether it would be useful or valid, it would make sense to see some practical examples that the public could examine to see if it is useful for them. Otherwise, we may wind up building something that we like, but not something the public likes.

-565

Share