Hi, everyone.
First of all, thanks very much for the positive feedback, and for signing at the page on Meta. That's a pretty stellar team already! :)
I'll respond to some of the comments here, but my aim is to keep the useful information up-to-date on Meta, so I'll be pasting the useful stuff back to Meta, and I encourage those of you who can deal with basic markup (surely at least most of you?) to continue refining the idea on the Meta page and to continue the conversation on its talk page.
On Sun, Aug 18, 2013 at 9:35 AM, Bob Kosovsky bobkosovsky@nypl.org wrote:
It's an interesting idea and suggests to me crowdsourcing subjects through something like tagging. For many years library staff have known that the manner in which subject headings are assigned for library materials is questionable, yet to come up with a better scheme is difficult. One problem that tagging shows is that people's command of language is very different. If there wasn't a controlled vocabulary or thesaurus, people would create numerous tags that, once combined with other articles, might be less useful than doing a full-text search.
Agreed. That's why I'm proposing using controlled vocabularies, alongside curated datasets such as Wikidata. That way we'll avoid having multiple unlinked variations on [[G. K. Chesterton]]'s name, for example.
It's interesting to me that databases like JSTOR don't use subject headings except with regard to the discipline of the journal where the article first appeared. They depend on relevancy rankings to assist users in finding articles.
Isn't that most likely because JSTOR don't have ready access to, or experience with engaging, a massive volunteer base who would undertake the work of classifying articles by subject headings? What I'm suggesting is obviously useful for JSTOR content as well, though I'd bet a good portion of JSTOR material is already topic-indexed fairly well in disciplinary bibliographical journals and databases (e.g. in my own academic field, classics, that would be L'Année Philologique[1]), so I'm leaving JSTOR out of the initial scope, for now.
Then there's the RILM database which established what it thought were fixed broad subject areas, but which are messy once interdisciplinary articles show up.
The way I'm approaching this is multiple, overlapping classifications, in multiple languages and according to multiple classification systems.[2] With some upvoting/downvoting or similar mechanism, I believe it can adequately solve interdisciplinary works.
Perhaps you can show an example of a single article with the kind of system
you're proposing?
Sure, let's see:
So, T.S. Eliot's "Hamlet and His Problems" -- http://en.wikisource.org/wiki/The_Sacred_Wood/Hamlet_and_His_Problems -- could be classified as ABOUT (or dc:subject[3], etc.): 1. http://lccn.loc.gov/sh85058566 -- "Hamlet (Legendary character)" (this is from the Library of Congress Subject Headings) 2. http://lccn.loc.gov/sh2008112835 -- "Theater--England--History--16th century" (likewise) 3. http://www.wikidata.org/wiki/Q2447542 -- "Prince Hamlet" (an item on Wikidata, about the fictional character Hamlet) -- sufficient to retrieve multi-lingual labels, link to Wikipedia articles, etc. 4. http://www.wikidata.org/wiki/Q41567 -- "Hamlet" (an item on Wikidata, about the play by Shakespeare) -- likewise 5-8. (a few more in English) 9-19. subject headings from some other thesaurus (do the DNB or BNF share their subject authority files like the LoC?)
All of these classifications are stored (either as Linked Data triples or in some conventional RDBMS [exposable as triples]) and can then be reviewed, revised, upvoted/downvoted, and of course searched.
I hope that is clear?
Cheers,
Asaf
P.S. I'm posting on this list with my Wikimedia Foundation address, because that's the one I'm subscribed to the list with, but let me clarify once again -- this is my own volunteer initiative, stemming from a longtime personal interest, and is neither endorsed by nor officially on the agenda of the Wikimedia Foundation.
[1] http://en.wikipedia.org/wiki/L%27Ann%C3%A9e_philologique [2] Do people still say "folksonomy"? ;) [3] http://dublincore.org/documents/2012/06/14/dcmi-terms/?v=elements#subject
On Sun, Aug 18, 2013 at 1:38 AM, Asaf Bartov abartov@wikimedia.orgwrote:
Hello.
Some of you have heard me rant about this for a couple of years now. So, I finally wrote something up:
https://meta.wikimedia.org/wiki/Massively-Multiplayer_Online_Bibliography
Much, much to be added, but I'd love for this to be a group conversation, so by all means, dig in! :)
A.