On 09/11/2012 00:56, Alexander Johannesen wrote:
Yup, have to agree with Jonathan here; primary sources are well and fine is as much as they can be hard to find when searching for “world war i”. However, I doubt anyone looking for “world war i” at this junction in their search even cares about primary sources. Or, put differently, every search is a journey, the search queries change depending on what they find and their level of investigation.
Can’t agree with that one. I think it is vital that for people who are interested in, e.g. the economic aspects of World War I, they need to be able to find Keynes’ “The Economic Consequences of the Peace”, the personal memoirs of what people lived through and on and on. If somebody did not put “World War I” (or in LCSH-speak “World War, 1914-1918”) nobody could find them. To assume that people don’t care about primary sources is too big of an assumption to make, in my opinion. Certainly if someone is not aware of something, that is a pretty good guarantee that they won’t want it, just like if somebody believes that when they click on the tag for Brazil in the Smithsonian project on Flickr that I discussed in the opening message of this thread, they are looking at the images of Brazil. What else could they think? Do the people who click on “Brazil” really not care one whit what they retrieve? “Whatever comes up is fine with me.”
Really? I can only hope the fellow working on my natural gas doesn’t have the same attitude, otherwise the entire block could go up. I think people accept things like the Brazil example only because they don’t understand what they are really looking at. They don’t know they are getting only two hits when there should be nine–and probably lots and lots more when you add in the entire set of Smithsonian images on Flickr. So if this is the reality, the task becomes something quite different: you can either let the public know the limits of what they are getting and I bet they would conclude that the tool you have made is pretty bad. Or, you can keep quiet, don’t let anybody know what they are really seeing, and hope to keep the random, limited result a secret. That is a real professional attitude!
The way I look at it, we either take the task seriously or we do not. If we don’t take the task seriously, then it will be difficult for others to take us seriously.
Yeah, I know, classical problems of library meta data collections. I say; put some more focus on that LCSH thing, because there’s gold in there when linked with something else, but you need to find it and use it better, and understand the problem of faceted exclusion in juxtaposition to open search. Or something.
This I agree with and I have never seen an implementation of LCSH that has really worked. That Trove site is very nice, but even better I think is another Australian project http://ll01.nla.gov.au/search.jsp which I think is the best I have seen so far. Your example of Puccini http://ll01.nla.gov.au/search.jsp?searchTerm=puccini%20giacomo shows off very well here, where the “Related Subjects” in the right menu can show either the complete headings or the facets, plus there are all kinds of other things.
Shirkey’s end statement at Educause was great: “Do not put together an interdisciplinary team from 12 departments and give them a budget of a quarter of a million dollars, and a year and a half deadline. Find five people and ask them what can you do in a month—for free. I think the results will surprise you.”