Posting to Autocat
On 9/6/2014 12:17 AM, Brian Briscoe wrote:
… speed has somehow become as, if not more, important than the quality of the results of a search. The quality of results has become secondary. Users are accustomed to a different way of searching, even though it is less precise and often returns poorer results. We must counter with a product that comes close to the same speed; is similar enough to the “Google box” that it is not repulsive; and will show the user the precision and quality that they are unaware they are not getting with keyword.
IMO, the way to do so, is not to bury our authority records in the way that I see so many ILS’ doing. I suspect we do so to look more like a keyword > search, but I think that strategy is the quickest way to make authorities unimportant to our users, directors and funders.
We have to find a way to steer the direction that our vendors are going.
I completely agree, but I will say that the traditional catalog–when people knew how to work with it as they were supposed to–actually could save the searcher’s time, but of course that didn’t mean it was perfect. OPACs never really have allowed these powers to be utilized. And that is why I have maintained that our current catalogs are broken and have been for a long, long time.
As a very simple example, let us suppose that I am interested in battles of WWII. I can search keyword “battles wwii” (what else would a normal person–scholar or not–search today?) and get lots of records to work with and may be “happy” enough.
But, our “information literacy” programs say that people are supposed to know how the catalog works and should therefore realize that this is a very poor search. They should know that the correct search is:
World War, 1939-1945–Aerial operations. (and/or)
World War, 1939-1945–Campaigns. (and/or)
World War, 1939-1945–Naval operations.
How are people to know this? Because of the directions found in the LC Catalog where the invaluable cross-reference “World War, 1939-1945–Battles, sieges, etc.” exists as a 450 cross-reference in each of these headings; e.g. to “World War, 1939-1945–Aerial operations”, and once there, you also find some handy references to other headings: Aeronautics, Military; Air warfare; Bombing, Aerial; Naval aviation, that probably would never enter into someone’s head. The same goes for “Campaigns” and “Naval operations”. As a result, this saves time when compared to a mere keyword search where you wind up looking at one record after another after another….
This was how “information search and retrieval” took place before computerization of full text and algorithms. It worked this way in other places than library catalogs: indexes found in the back of books worked this way, the yellow pages in the phone books and lots of other places, although the structures there were much less extensive. Of course, it wasn’t perfect, but that is how the tools were (and still are) designed to work. You take away the cross-references and handy notes, of course people will flail around.
The very few times I have managed to get a non-cataloger to understand how the system is supposed to work, they suddenly become much, much less enamored of those full-text, algorithmic search results and in fact, they have asked why these wonderful cross-references are so hard to find in library catalogs, and why the general search tools such as Google don’t implement these methods that actually could–and they see that it should–save their time! Good questions and I have no answers!
How are people today supposed to find the cross-reference for wwii battles? By doing something very strange: by searching for subject (!), and entering exactly: “world war 1939-1945–battles”. The chances of that happening probably come close to the proverbial roomful of monkeys randomly typing out the complete works of Shakespeare!
I still maintain that people would actually like and use our tools IF they were reconfigured to work for the 21st century and not for the 19th century. Until we admit that our catalogs are broken in fundamental ways and must be fixed (because they CAN be fixed), any changes to our rules (adding relator codes and spelling out abbreviations), adding FRBR relationships, changing formats, and adding linked data until it’s coming out of ears, will end up making zero difference to the public until the catalogs themselves are fixed. We can do all of those things and our catalogs will still be broken. If the only way people can get into our authority structures is by doing left-anchored text string searches, then our authority records are fated to forever be irrelevant.
But catalogers CANNOT fix this because the problem is not with the rules, the formats or anything that catalogers can work with. The problem is in the “innards” of the catalog and it is the systems people who must fix it so that our authority records can actually be found in ways people search today in the 21st century, and not pretend that there is no real problems because people can just search by left-anchored text strings. Nobody does that. Even then it doesn’t work–not even for me–as I demonstrated in (I am ever shameless!) Cataloging Matters no. 18: Problems with Library Catalogs It just DOESN’T WORK! That means, it is broken–not broken beyond repair because it can be fixed–and until this is accepted there will be no attempts toward fixing it.
I don’t know what would work, although I have some ideas that could be shown right or wrong in the end. To figure out what would work would demand study, thought, consideration and reconsideration as to what would be the most useful for the public today. And not only should catalogers be involved, but the entire community. The catalog is for them and not just for us.
I can only hope that these reconsiderations will happen eventually.