Christina's LIS Rant
Friday, April 01, 2005
O'Reilly Emerging Technology Conference 2005: Shirky, "Ontology is Overrated"
CS acts like he's saying something new
Argh. Yes, as a matter of fact, we all learned this in school. LC developed to provide access to a particular collection of books: the books in the physical collection of the library of congress. It was never meant to be a representation of the state of the world's knowledge.
CS also gives the example of the lack of a category for oncology. Right, that's why NLM has their own classification and MeSH.
DDC has a lot of problems, too. Too many computer subjects not originally accounted for by Melvil, etc. That's why there are huge committees and people trying to fix and update these things. The end goal is the same: to provide access to library resources to solve the information needs of the end user. Library classification systems cannot solve the philosophical questions that should be left to PhDs and theologians.
Update: The link to the mp3 of the presentation was provided in the comments to this post. I'll include it here for everyone's benefit: http://www.itconversations.com/shows/detail470.html . Perhaps I was too hasty in my post above. From the more detailed description on this site, it sounds a lot more interesting.
Update 2: I like this version of social tagging.
Update 3: Listening to it now. Does anonymous know where I can get the slides? Still a little annoyed at some of the discussion of lc stuff, but I guess he needs to go about it that way due to his audience. "signal loss" in traditional classification schemes where you can have overlap etc., in online tags...
Update 4 (5/17/2005): Shirky has posted notes from this and an IMCExpo talk here. ¶ 9:00 AM| |cites (technorati) |
the Library of Congress's classification scheme exists not because concepts require consistent hierarchical placement, but because books do
Argh. Yes, as a matter of fact, we all learned this in school. LC developed to provide access to a particular collection of books: the books in the physical collection of the library of congress. It was never meant to be a representation of the state of the world's knowledge.
CS also gives the example of the lack of a category for oncology. Right, that's why NLM has their own classification and MeSH.
DDC has a lot of problems, too. Too many computer subjects not originally accounted for by Melvil, etc. That's why there are huge committees and people trying to fix and update these things. The end goal is the same: to provide access to library resources to solve the information needs of the end user. Library classification systems cannot solve the philosophical questions that should be left to PhDs and theologians.
Update: The link to the mp3 of the presentation was provided in the comments to this post. I'll include it here for everyone's benefit: http://www.itconversations.com/shows/detail470.html . Perhaps I was too hasty in my post above. From the more detailed description on this site, it sounds a lot more interesting.
Update 2: I like this version of social tagging.
Update 3: Listening to it now. Does anonymous know where I can get the slides? Still a little annoyed at some of the discussion of lc stuff, but I guess he needs to go about it that way due to his audience. "signal loss" in traditional classification schemes where you can have overlap etc., in online tags...
Update 4 (5/17/2005): Shirky has posted notes from this and an IMCExpo talk here. ¶ 9:00 AM| |cites (technorati) |
Comments:
My reaction is also...ummm...is Clay aware that LC Classification and LCSH are two distinct tools?
I'd like to hear some intelligent arguments about social tagging as an alternative to a controlled vocabulary taxonomy (and personally think that a controlled vocabulary system can be built using social tools--but that gushing on about flickr and del.icio.us totally ignores the benefit of precise language and preferred terms).
My ideal: an LCSH-type vocabulary built using wiki-type tools, with the same editorial give-and-take as seen in Wikipedia.
I'd like to hear some intelligent arguments about social tagging as an alternative to a controlled vocabulary taxonomy (and personally think that a controlled vocabulary system can be built using social tools--but that gushing on about flickr and del.icio.us totally ignores the benefit of precise language and preferred terms).
My ideal: an LCSH-type vocabulary built using wiki-type tools, with the same editorial give-and-take as seen in Wikipedia.
# posted by 
: 4/06/2005 9:45 PM
: 4/06/2005 9:45 PM
I'd like to see multiple competing thesauri -- like PACS and MeSH... so you pick a term and pick a system from which it comes. The fact is that no one list of controlled vocabulary fits all but having ZERO control of the vocabulary fails miserably, too.
A friend of mine made a comment in a totally unrelated discussion to this one about ontology, but relevant here. He was speaking about classifications of western and non-western music, and the "epistemology" behind why we define categories the way we do:
"I suppose that the epistemology of this issue is partly what underlies the intensity of discussion between [the participants]. It's important to know what it is we know ... or at least to be able to describe what it is we're talking about, and what it isn't. Even if the edges are fuzzy. But once one draws a boundary around a given community of speakers, the issue is no longer about the arbitrariness of symbols, but about what produces meaning for the users of those symbols."
Post a Comment
"I suppose that the epistemology of this issue is partly what underlies the intensity of discussion between [the participants]. It's important to know what it is we know ... or at least to be able to describe what it is we're talking about, and what it isn't. Even if the edges are fuzzy. But once one draws a boundary around a given community of speakers, the issue is no longer about the arbitrariness of symbols, but about what produces meaning for the users of those symbols."
