Are LibraryThing tags mediated or controlled?
Join LibraryThing to post.
This topic is currently marked as "dormant"—the last message is more than 90 days old. You can revive it by posting a reply.
I'm a library student doing some analysis of LibraryThing tags, and while looking through a number of tag clouds, something occurred to me: I'm not seeing any obviously misspelled tags, ie somebody tagging a book as "sicence" rather than "science." Clearly tags can be quite individual, but with so many people tagging, you would think you'd see quite a lot of that.
So my question is, does LibraryThing mediate or control its tags to weed out obvious mistakes? And if so, does it control for other qualities as well?
Any insight would be greatly appreciated; I've been digging through the site and all these conversations, but I haven't found an answer to this yet. :)
Any user can propose to combine two tags, and when combined only the most popular of the combined tags will show up in tag clouds. So, to use your example, the tag "sicence" has been combined into the tag "science". Since the correctly spelling is used much more often, that is the variation that shows up in tag clouds.
To expand a bit on what norabelle414 said, and glossing over the specific details of combinations (which require a supermajority of those who choose to vote to actually take effect):
You can still see all individual tags that have been combined into the "majority" tag on the tag page, like this:
Up at the top there's a line reading "Includes", which lists all the tags that have been combined into the main tag. Click on any of them, and you can drill down and see the users who employ that variant, and the books they tag, for instance:
This allows you to examine the usage of slight variants, translations, and typos without losing the social benefits of viewing synonymous tags together.
Right. The second effect is that, even if "sicence" wasn't combined, it probably wouldn't be among the top tags for any item. So, for anything even moderately popular, you'd never see it unless you hit "show all tags." But, as Nora explained, tag combination does a lot of the work here.
So any control is performed by users?
Awesome, thanks y'all! I appreciate it. :)
As another wrinkle our "LibraryThing for Libraries" product does use some tag selection by professionals.
Yeah, I've been looking at that as well... but as this is just supposed to be a 20-page paper, and not my doctoral thesis (at least, not yet) I'm just going to ignore that for the time being. Folksonomies and controlled vocabularies are kind of a bottomless well... :)
They are indeed!
Incidentally, I recently changed the algorithm slightly. It still picks the most common tag among all variants, but it balances the most common by number against the most common by user. Basically, one member who tags a LOT can't sway the top tag too much. The net effect is to dethrone a few Dutch and German words as the top version for some rare tags that didn't get that much play.
Please send questions, thoughts, etc. I can talk about tags ALL DAY!
A video I did about LT in general, with a lot on tags: http://vimeo.com/7953189
#6: Obscenity filtering is an interesting question. I think looking only at tags that multiple people have applied to a book cuts out most of the gratuitous use. The main exception that jumped out in fuck was that fuck was applied to Catcher in the Rye three times, which would likely be filtered out anyway for such a popular book. The F-Word and Fuck this Book are natural places for the fuck tag. Zipless fuck is used repeatedly on a book that talks about the zipless fuck. Mindfuck is used a lot; unlike the last two, I suppose I can see why people wouldn't want it used on a library tag list, as the obscenity isn't essential to the book. Cunt is used on three books multiple times, two about the word and one book of pictures of genitalia, where I find it inappropriate in a public sense, but then again I haven't read the book. (I think I understand and appreciate why it was used; I'm not objecting to it in LibraryThing; I wouldn't want it so attached in a curated collection.) Nigger is used repeatedly in a book about the word and Huckleberry Finn.
Naturally, some of the cases are context sensitive and impossible for a computer to distinguish. The F-Word and Cunt: A Declaration of Independence should come under their tag, and that's a little important, because the F-Word can't be found by a search for Fuck if not for tags or subjects.
For LibraryThing for Libraries we have an language filter libraries can turn on. Basically, a few tags have ratings—PG, R, NC-17. If you don't want the "oral sex" tag to show up, you can turn on the filter and set it low. Some libraries use it, some don't. It was a lot of fun coming up with the list of words.
What type of libraries have disabled the "oral sex" tag? I can't imagine why a library that would carry She Comes First: The Thinking Man's Guide to Pleasuring a Woman would drop the oral sex tag, but I can see why one that carried Sandpiper might. What about the other tags on Sandpiper: promiscuity, rape, sex, sexuality, sexual harassment, sexual intimacy, sexual situations, teen sex and teen sexuality?
12 -- You have interesting reading habits. I can see you've given this a lot of thought.
If you don't want the "oral sex" tag to show up, you can turn on the filter and set it low.
I suspect that "off" might be a better setting than "high" -- unless you're fairly certain nobody will be coming through with, say, a giraffe.
Just getting back to this, but thank you for the invitation! I've read a few bits you've written around here, but as I dig down into the nitty gritty, I might very well have some technical questions.
I'm really interested in how tagging and folksonomies are ultimately going to fit into library catalogs and OPACs, and this is hopefully just my first exploration of the subject.
Yeah, but are they? For all the hoopla, LTFL is the only tagging system that's taken off in libraries, and it came from outside libraries. There are dozens of systems now that allow tagging, and have maybe 1,000 tags. It's lead to something of a backlash against tags—justified in some sense and unjustified because the companies that sell tagging as a "feature," sold social software without creating or setting the groundwork for a "society" that makes it.
This group does not accept members.
This topic is not marked as primarily about any work, author or other topic.