Wikipedia has a way of handling the Paris example -- look at the link just under the article title, and you will see 'This article is about the capital of France. For other uses, see Paris (disambiguation) [link]', which takes you to this page:
http://en.wikipedia.org/wiki/Paris_%28disambiguation%29
With alphabetical ordering inside of designated meaningful areas. And it finds Paris, Kentucky.
Celeste
-----Original Message-----
From: Next generation catalogs for libraries [mailto:NGC4LIB_at_LISTSERV.ND.EDU] On Behalf Of Karen Coyle
Sent: Monday, November 12, 2012 12:15 PM
To: NGC4LIB_at_LISTSERV.ND.EDU
Subject: Re: [NGC4LIB] Video of "Think Different"
On 11/11/12 1:58 PM, Dan Matei wrote:
> My Paris example is a bad one :-( My point is: "lexical co-location"
> is could be valuable for the user. Suppose I'm looking for
> "Stockhausen" (knowing exactly that I mean Karlheinz S.). But if the
> name "Markus Stockhausen" is shown to me after K.S., maybe this way I
> find out that Karlheinz has a son, composer too. Dan
Dan,
But if there was a daughter Mabel who marries Mr. Smith and becomes Mabel Smith... :-)
What you have working for you here is serendipity, and even your Paris example could work sometimes because of the serendipity. (How many of us find ourselves reading nearby words when we've done a dictionary look-up, just because we can?) However, using my earlier criterion: is this the best solution for the user? I'd say that what you've described here seems to me to be a role for a "see also" relationship -- which then wouldn't rely on the accident of language. So your example provides some serendipity, but if your intention is to get users to "nearby persons" then some other mechanism will be needed. I worry that alphabetical order may be deceptive in a fair number of cases because it looks so... complete.
Which leads me to something else that I have been thinking about (and that I presented briefly in my talk), which is the way that alphabetic collocation brings together things that have nothing to do with each other, and separates related concepts. So you get alphabetical runs that go like:
ball
ball-bearing
ball boy
ballet
ballistic
balloon
ballot
ballroom
and you've gone through a whole lot of different concepts, from sports to classical dance, military themes, toys (or weather), voting and architecture. Where terms or roots have been combined (ball and
ball-bearing) there might be something that would help a searcher, but the connection between "ball" and "ballroom" has probably been lost because the intervening terms are unrelated to one of the meanings of the root "ball" (an event) and the user will probably give up before reaching "ballroom." (Oh, and then there is the question of spaces and punctuation -- I love to trot out the ALA filing rules for non-librarian friends. They get a good laugh out of it. I could cry.)
Many of these problems of using terms out of context (and alphabetical order is out of context) also pertain to keyword searching, which I think works primarily for proper nouns or for clearly named concepts ("solar energy"). As I've said before, there are many concepts that are hard to search on because they don't have clear, agreed on names, or because the term has multiple meanings (python, library, pluto). And there is no way to search on a concept that you cannot name.
*sigh*
kc
--
Karen Coyle
kcoyle@kcoyle.net http://kcoyle.net
ph: 1-510-540-7596
m: 1-510-435-8234
skype: kcoylenet
Received on Sun Nov 11 2012 - 19:09:11 EST