Cataloging and fun
On Thursday we introduced a silly new "meme" page called "Dead or Alive?" which listed your authors by their mortal status--alive, dead, unknown or "not a person." (See the blog post or check out yours.) The feature drew on the birth and death dates of the authors in our Common Knowledge system, a free (Creative Commons) "fielded wiki" for miscellaneous "cataloging" information (think "Wikipedia for book info"). To move an author from the "unknown" column, members had to find their dates and enter them onto into Common Knowledge.
Here's a chart of Common Knowledge contributions over the last month.* Can you spot the day "Dead or Alive?" went live?
As you can see, birth, death and gender edits (gender is where you mark an author as "not a person") went through the roof when the feature was announced—from an average of 143 edits per day, to 3731 and 3584, 25 times the average. Other edits went up too—a 30% increase.
A few members joked that it was a plot to encourage contributions to Common Knowledge. It wasn't that. I just thought it was a funny idea, but I wasn't unaware that it would have that effect. Indeed, the upshot shows again something of a LibraryThing finding—that regular people will contribute cataloging information if you make it meaningful to them. That is, whatever incentive there is to add author information, the incentive is increased when they're your authors, and increased again when that information does something for you. Of course, even if incentive is personal, the effect is general; you update the author because you have his or her book, but everyone else shares in the value of that update.
The way this works undercuts a common myth of "Web 2.0"—that there are all these people out there adding "user-generated content" out of altruism or an extreme mismatch between time and exciting things to do. And it cuts against an older myth, that cataloging is so boring you have to pay people to do it.
We've seen the same jump every time we introduced a new Common Knowledge category, and again when we made that category "come alive" in some way for members. And although the short-term jump will surely level out, the overall rate of "dead-or-alive" entries certainly not. You get more changes when the changes do something for people.
Now, of course, there's a whole list of things this doesn't mean. It doesn't mean that LibraryThing members are doing their job well (although I suspect they are). It doesn't mean the same would apply to much more difficult forms of cataloging, or to forms that generally presuppose professional training (ie., LCSH). And it doesn't mean that regular people will get to the "rare stuff," indeed it probably means that average cataloging attention is directly related to popularity of the underlying item.
Even so, pretty cool. Oh, by the way, I'm adding a feature allowing you to compare yourself to other members, which should inflame the other great motive for personal metadata—competition. After all, my library has a higher dead/alive ratio than yours!
UPDATE: Here's the current chart, without day-norming. Notice how everything went up.
*The numbers are normed against day-related changes. Basically, we smoothed out that many more edits are made on Monday than Saturday.
Here's a chart of Common Knowledge contributions over the last month.* Can you spot the day "Dead or Alive?" went live?
As you can see, birth, death and gender edits (gender is where you mark an author as "not a person") went through the roof when the feature was announced—from an average of 143 edits per day, to 3731 and 3584, 25 times the average. Other edits went up too—a 30% increase.
A few members joked that it was a plot to encourage contributions to Common Knowledge. It wasn't that. I just thought it was a funny idea, but I wasn't unaware that it would have that effect. Indeed, the upshot shows again something of a LibraryThing finding—that regular people will contribute cataloging information if you make it meaningful to them. That is, whatever incentive there is to add author information, the incentive is increased when they're your authors, and increased again when that information does something for you. Of course, even if incentive is personal, the effect is general; you update the author because you have his or her book, but everyone else shares in the value of that update.
The way this works undercuts a common myth of "Web 2.0"—that there are all these people out there adding "user-generated content" out of altruism or an extreme mismatch between time and exciting things to do. And it cuts against an older myth, that cataloging is so boring you have to pay people to do it.
We've seen the same jump every time we introduced a new Common Knowledge category, and again when we made that category "come alive" in some way for members. And although the short-term jump will surely level out, the overall rate of "dead-or-alive" entries certainly not. You get more changes when the changes do something for people.
Now, of course, there's a whole list of things this doesn't mean. It doesn't mean that LibraryThing members are doing their job well (although I suspect they are). It doesn't mean the same would apply to much more difficult forms of cataloging, or to forms that generally presuppose professional training (ie., LCSH). And it doesn't mean that regular people will get to the "rare stuff," indeed it probably means that average cataloging attention is directly related to popularity of the underlying item.
Even so, pretty cool. Oh, by the way, I'm adding a feature allowing you to compare yourself to other members, which should inflame the other great motive for personal metadata—competition. After all, my library has a higher dead/alive ratio than yours!
UPDATE: Here's the current chart, without day-norming. Notice how everything went up.
*The numbers are normed against day-related changes. Basically, we smoothed out that many more edits are made on Monday than Saturday.
Labels: cataloging, dead or alive, Social Cataloging, zombies
posted by Tim at 3:23 PM
13 Comments:
I think you should consider this sort of effect next time you claim that there's no point in making the date fields do anything because no one uses them :P
"which should inflame the other great motive for personal metadata—competition. After all, my library has a higher dead/alive ratio than yours!"
I never quite recovered from the humiliation of having my obscurity value go above 1. Imagine how embarrassing it would be if I turned out to have more living authors than dead. *shudder*
the upshot shows again something of a LibraryThing finding—that regular people will contribute cataloging information if you make it meaningful to them
Dates! Dates! Make the dates do something!
....Oh, I see Zoe's already beaten me to the punch. But still... dates!
The jump in birth/death/gender is fabulous, but I also see quite a jump in other CK contributions compared to the same days of previous weeks. Once you're in there editing CK, it's a lot easier to tweak a canonical name that never wants to display right, or add the corresponding places when you've already looked up the dates.
And it doesn't mean that regular people will get to the "rare stuff,"
Though of course the handful I entered were moderately rare -- poets and so forth whom I am fond of (but who are popular enough to have Wiki pages with the relevant info). So that level of rare. Because I really wanted that info to be there.
Wow, social networking at it's best.
I'm not sure about the calculations though
"Dead: 11 / Alive: 41 / Unknown: 7 / Not a Person: 0
Percent dead: 78.85%"
Shouldn't that be "percent alive"?
:)
Who'd have thought that you'd get people interested in entering DATES when they turn out to do something interesting? Of course started DATES, finished DATES, are hardly ever used, so there's no point in letting cD play around a bit more with lttl ... or is there?
Why, you could even do something like:
Library movements in the last < 0|30|60|90 > days
Acquired
Started
Finished
Offered
Disposed
I mentioned this on the original blog post, but I think a similar page showing gender split of your authors would be a neat feature. And have the side benefit of increasing data population.
Great to see these spikes! Tim makes good points here re Web 2.0.
what, people don't use the date fields? huh. but, but... you can line them up and presto, a chronological list of every book you ever purchased. and you can also arrange them into a list of all the books you've ever read, in the order you've read them. after all, what you read in the 60s might differ substantially from what you're choosing to read today. okay, not all of that is socially significant maybe, but on a personal level that kind of detail is surely a really fine thing. i'd like a spot where i can input where i bought them (both place and bookshop), or who gave them to me as a gift: sometimes that's very personal, and sometimes very social, but it's part of why we love them: every copy of every book we own has a personal history, and to my mind we're not documenting yet enough of that. and who knows, maybe you wandered into the Strand ten years later and bought that same book to cherish too, which creates a kind of bond between us across time, out of place.
and surely how long you've kept a book (or re-bought it, maybe more than once, because you couldn't live without it: there's another little thing that might be nice) has some social meaning. a book that's survived every cut, move, change of interests, and personal upheaval over a forty year period, for instance, speaks more eloquently to the class of Books That Matter than does the book we all bought, skimmed, and recycled only yesterday. books contain memories, ghosts, that can be captured by the existence of that kind of data. and bookstores too, you know the kind, might find it interesting to see who bought their books to cherish. that tree didn't die just to support their bottom line!
and speaking of that, how about a checkmark to distinguish between what we own, what we used to own, and what we want to own someday; and between books we read that were ours, as opposed to books not-ours but read on the net or borrowed from a friend or a library. because all of that surely has some social meaning too. it's an invisible history of what the book means to us, all that data buried in the history of how long we've kept it. a very personal thing, bien sur, but surely a part of why we catalog them too. plus, wouldn't it be kind of cool to know who in the LT db has owned that particular book the longest? or how many people have discarded their copies over time? or just what bookstores (a lot of which are in this db already as locations) we cite?
As Raven said, the % seems to be backwards.
Dead: 28 / Alive: 85 / Unknown: 104 / Not a Person: 3
Percent dead: 75.22%
(And yes, I've been working on cleaning up my unknowns.)
I've been adding tons of stuff to CK that I find when I look for an author's birth/death dates. If I get it off a decent bio, I can add various residences, education, occupations, and such. I figure as long as I'm already there . . .
Thanks so much for the time sink, Tim!
Death, life—all part of the same process.
(bug fixed)
Couldn't you do a male/female meme as well?
Post a Comment
<< Home