Este tópico está presentemente marcado como "adormecido"—a última mensagem tem mais de 90 dias. Pode acordar o tópico publicando uma resposta.
There's a great pride in having eclectic tastes, but in my case, the obscurity is a function of having lived in Australia and Japan, and buying locally published titles.
I actually own over 2000 books and started cataloging my books in another program. But it doesn't work well without an ISBN. I believe it's because it's not linked to as many sources as Library Thing.
Do you think there is some meaning to having an obscure reading list? What does that say about someone?
As the result of his Language Log post, I discovered that there actually is at least one user whose weighted list I show up on (per the group description), though down in the middle somewhere. I suspect that's just an accident of the fraction of his library he has entered so far: some foreign language grammars and pre-Graham LISP references. It should vanish back off once he gets a chance to enter more. What does waiting for that say about someone?
9battlinjack Primeira Mensagem
One of the things that I've been nagging Tim & co. for is some sort of "obscurity chart" that would let folks know where they stood in relation to the rest of the L.T. community.
As I understand it, the various stats things are rather broken at the moment (for example, the page with the data about reviews has not updated since September), and the whole issue of the "Zeitgeist" is under consideration, but given a lower priority than fixing the search function and streamlining the language issues.
Fogies, my most popular is Life of Pi (Moby Dick was a library book) and #100 is Iran awakening. But this is just the result of my slow progress. 10% is probably a better measure - my 10% down book (at least, today) is Much Depends on Dinner.
Is that for median or mean?
-- Bob Campbell
It really bugs me when I find I'm reading a "popular" book, knowing that it's going to trash my obscurity rating ... I've even made reading decisions based on not wanting to add in books that were "top 100" ... if the obscurity rating were based on percentage of exclusivity, it would be resistant to having any particular book mess up the numbers!
Nope. "Percent unshared" doesn't work, because that statistic privileges works in less-common languages.
(E.g., If someone has collection here that consists only of works in Lithuanian (say), and there are no other LT members with those books...).
"Percentage unshared" is an interesting problem - my statistic has been dropping steadily as
a) I clean up my own data
b) elves* come in and clean up the shared author/works data
c) new members who hold previously-unique works join up.
I was up in the 15-20% "unshared" ballpark,
but that's been steadily dropping until I'm now down to 7% unshared. (My "obscurity rating" is now 17/101, up from around 8/70 six months ago.)
*Well, I've seen my own data get tidied up overnight, and without my participation. "Elves" is as good a description of this effect as any.
That is obscure, mine is now 5/112 with 1725 books cataloged. We actually share one book, Flowers; a guide to familiar American wildflowers by Herbert Spencer Zim. In the last three days, I' have been catloging some of my wife's plant books, which all seem to have few sharers. So, maybe these types of books are uncommon for the typical librarything.com person. My number went up because I was also cataloging my wife's classsics, Shakespeare, Hemingway, Maugham, etc.
-- Bob Campbell
Many people don't put their whole library in librarything.com, so they may be skewing their numbers by what they leave out. (i.e. those 10 32-page that my grandmother gave me aren't worth anything, so I won't put them in. But these are obscure items in 2006 even if they were common in 1930).
We have an obscurity of 7/362 (we blame Harry Potter, too, for the latter figure) on a library of 3,492 ... But that is all the ISBN's books ... the non-ISBNs are going to push the obscurity still further, methinks...
You are so right about how fascinating obscurity is! While I can see other things are (rightfully) prioritised above obscurity getting into zeitgeist or some other comparative function, I still can't wait!
Pat from Rivendell
See, this is the sort of thing which makes me ache for an alternate system for measuring obscurity. In the curent calculations, you can be seriously impacted by adding one super-popular book, and I'm sure that many of us who value "obscurity" as an attribute have self-edited to avoid this happening ... I know that I've steered away from particular books which I had a thought to read because I knew it would have a deleterious effect on my obscurity rating!
I've previously suggested a system for measuring obscurity where 1/# was the core measurement ... with those added (1/#a + 1/#b + 1/#c, etc.) and divided by the total number to render a fraction which was close to 1 for an extremely obscure library, and infinitesimally small for an extremely common library. This was thrashed out several months ago in the old Google Group, but nothing ever came of it.
I would be interested to know what the median obscurity score of all LibraryThing users is.
I just looked up the median obscurity scores of the first 40 people under "recent users", and this is what I got:
So it looks like the median median book obscurity is somewhere around 200, and that if you want to be considered in the top 10% of obscure people you should perhaps have a median book obscurity of no higher than 15.
My obscurity is 110/590, so I won't be joining this group :)
My obscurity keeps decaying ... I'm now at 15/240 ... which by this figuring would leave me teetering on the edge of "insufficient obscurity"! This has happened very quickly (on 12/31/06 I was at 12/189), and it's not like I spent all of January reading Harry Potter or anything.
I find that books older than 50 years (from original publication) take longer to catalog, if thye are still in print. I've gotten over 100 hits on Amazon.com on some titles. Sometimes, I just put my old books in manually to save time wading through the librarything responses.
You may have caught LT when they were doing maintenance.
I agree about the lack of the added great libraries not being represented. Of the three great U.S. libraries, only the Library of Congress is represented, but that is a godsend when you have a an LC card number, but no ISBN (perhaps 8% of my library). The other two missing libraries in the U.S. are Harvard University and New York Public. However, Yale is there, so I guess I can be 2/3 satisfied on that.
It would be nice to have the British Library, also Oxford and Cambridge. Also, one or two great French libraries would be nice (Biblioteque nationale and Sorbonne - I don't really know French libraries). I do use the University of Montreal when I run dry on English sources.
- Bob Campbell
I, too, wish that Tim would add a "distribution graph" (like those for dates) on the Fun Statistics page that would show how one's obscurity numbers fall versus the LT population as a whole.
# 41: I think a distribution graph would be great.
1 - user: jhruby7
Compare the first and last five scores here with the ones I posted in February! Based on this the median median book obscurity score is now around 185-190. If you want to be in the top 10% of obscure people you should have a median obscurity score no higher than 4! This is the opposite of what I had expected.
It might be just due to randomness. On the other hand, I noticed that the users with the lowest median book obscurity scores tended to have small libraries (around 50 books) on specific topics. Perhaps for some reason LibraryThing has recently been attracting a higher proportion of "specialists" to "generalists" than when it first started, although I can't think why that would be.
EDIT: I just noticed that jhruby7 has the "want" tag on every single one of his 65 obscure books. So it's a wish list, not a list of books that he owns.
I still do wish that Tim & Co. would provide a graph that would show "where on the bell curve" a particular obscurity rating would fall!!
I own quite a lot of Dutch books which are certainly obscure and terribly hard to catalogue. I've manually entered nearly 100 already. Why isn't the Dutch Royal Library (Koninklijke Bibliotheek) on here? THe only Dutch cataloguing source is the Technical Univeristy in Delft which just doesn't have that extensive a library.
I've noticed that Dutch users of LT manually enter almost all their books, which must be why there are so few Dutch users.
Plus there is an article in the LT blog about what LT is doing in the Netherlands. It makes a reference to a more extensive comment at the "Thingology Post".
Tuesday, August 07, 2007
Will the Dutch-ness ever stop?
We just inked a deal with AquaBrowser, an Amsterdam-based library catalog company. See the Thingology Post.
posted by Tim at 8:12 AM 3 comments
Beyond this, there is a Dutch version of LT at:
Dutch site (www.librarything.nl)
Books cataloged: 42,807
Groups: 12 (see groups)
48--99% of my Dutch books don't exist in English translation so there's nothing to combine them with. I'm obscure in every way. Probably even the Dutch Librarything site doesn't include most of my books.
Now, with virtually everything on (there is some Japanese material that Mr Rivendell is still working on): 5/573.
When this thread started (at the end of last year), it was 12/189 ... the last time I posted about it (a scant 3.5 months ago), it was at 25/372 ... and now it's all the way up to 31/438!
I feel so irretrievably mundane!
I've been assured that this is a function of so many new people joining L.T. and bringing in books that I used to have exclusively (or at least obscurely), but that doesn't help me with feeling like I'm becoming normal or something hideous like that.
1,876 Number of distinct works*
1/3 Median/mean book obscurity**
I know my books are not 'obscure' so why the low figures?
If others own the same book as you but the copies are not listed together, you can combine them into works with other peoples copies which would increase your numbers.
Hope that made some kind of sense.
I noticed that a large proportion of your books you added by manual entry. Presumably you couldn't do a match with Amazon or LOC - this in itself would make your library more obscure as most of can make a match for the majority of our books. Your library looks like a specialized collection. If I had just put in my transportation and travel books, my obscurity would dip quite a bit.
As of today, I'm at 40/555 ... you would think that my not reading any fiction would make for more obscure ratings, but I keep slipping further into the "non-obscure".
Or maybe LT is adding more and more people so that the threshold for "obscure" is rising. You still have an obscure library, but a greater absolute number of people will share that with you even if your relative obscurity has unchanged.
It seems unlikely that LT is attracting a greater proportion of people with obscure books than it used to.
My figures long ago slipped into the mundane 398/1,398 although I don't own many of the most common LT works.
(Some day we will add our holdings in Chinese, Japanese and Korean. Look for 1/1 when that happens...)
Fogies, your rating has gone to 28/284 without your adding anything more. Just shows how LT has grown. I do hope you start on your East Asian books.
I feel so "mundane"!
I still don't fully understand the numbers, despite their meaning's being posted repeatedly.
If it will help, if you compare my number, 16/339 with that of Fogies 26/259, you can see that his two numbers are closer together. What this means is that I have a greater number of really obscure books (singletons), but also a bigger collection of really popular books (e.g. Harry Potter) and Fogies is somewhat more in a mid-section (but very broad). The second number is very influenced by very popular books, whereas the first number is more influenced by ones uncommon books.
Someday, when it becomes possible to get reference to the exact edition in, like, the Japanese Diet Library or the U. Peking library or some such, we'll see what can be done. We would love to have them cataloged, but already done. And also there's the problem that only a few hundred of the most recent have ISBN's.
(As an aside, and only in Boston, there once was a street which had NB & SB traffic on both sides of the median, the east side of the median was for slow traffic so that people could look at flowers in the park, and the west side of the median was for the express traffic).
12/171 Median/mean book obscurity**
** number of users who have books in your collection
So what does this mean is more layman's terms??? And how does it relate to the rest of LT? Does it depend on how many books you have that no one else has, or doesn't that matter?
The second number is simply the total of all your shared connections divided by your total book count. To illustrate on a simple example, only 3 books, one shared by 1 person, the second by 10, the third by 100; the average is 37.
As for what your numbers mean in the context of LT, for a proper answer we'd need an analysis of total LT member obscurity data. That we don't have, but you can get some sense of where you fall by checking out obscurity numbers of say... well, thirty randomly selected members with libraries of size comparable to yours. Far from perfect, but we can't do better than an approximation.
And, IMO, right now your numbers are fairly obscure. (But what does it MEAN? :))
As a combiner I've found this group to be a gold mine for orphaned editions that need to be combined. Sometimes there's not much I can do -- when the people crowing about their obscurity just left the author off of a popular work, for instance -- but I take secret pride in watching obscurity ratings creep up when they're the result of uncombined editions rather than actual obscurity.
I see the last time I posted a number was back in August when I was 45/603 ... today I added a book (#1,800 in my LT library) and found that my obscurity had further slipped to 53/663 (two years ago it was 12/189 ... of course, at that point LT was barely 15 months old, so had a lot less members/books).
I do wish Tim & Co. would do some charting for this ... it would be interesting to have graphs to put this stuff in context.
Here are our numbers from the beginning:
April, 2006: 1/2
June, 2006: 1/4
July 2006: 2/19
September, 2006: 2/38
October, 2006: 3/65
January, 2007: 4/105
March, 2007: 5/144
(Most of those changes I attribute to our adding more books. The jump between October 2006 and January 2007 for example includes our adding all of our Harry Potter books.)
March, 2008: 12/310
(Most of that change I attribute to the growth of LT as a whole. We added only a couple hundred books during that span and they were middling in popularity.)
Today January, 2009: 18/391
We've added another thousand or so books since March. I would have guessed that they were on average "more obscure" and would have worked against the upward pressure on our obscurity numbers, but so it goes.
As of 02/07/2011 I'm at 77/940 ... just over 2 years ago (the last time I noted the numbers) I was at 53/663 ... I started out at 3/36
Any others see the same pattern? (I think #46 is on the same track.)
73/672 Median/mean book obscurity
As of January 16, 2012.
I wish I had kept track of earlier numbers.
As of 01/15/2012 - 85/1,027 ... which is quite a leap from my last checking, somewhat less than a year ago (77/940).
Apparently a low median = lots of unshared books, but high mean = lots of books shared by lots of people.
Help, please? :)
John and Jane both own three books. They put them on LibraryThing, and sort their book collections by "total members".
In John's library, book #1 is owned by 100 people, book #2 is owned by 100 people, and book #3 is owned by 100 people.
John's mean book obscurity is the sum of members who own each book in John's library divided by the number of books John owns:
(100 + 100 + 100) / 3 = 300 / 3 = 100 mean book obscurity
John's median book obscurity is the number of people who own John's "middle" book when John's library is sorted by total members, in this case book #2:
100 median book obscurity
John's median/mean book obscurity is 100/100 (in this case the slash is just a separator, not a division operator).
In Jane's library, book #1 is owned by 1 person (Jane), book #2 is owned by 10 people, and book #3 is owned by 289 people.
Jane's mean book obscurity is the sum of members who own each book in Jane's library divided by the number of books Jane owns:
(1+ 10+ 289) / 3 = 300 / 3 = 100 mean book obscurity
Jane's median book obscurity is the number of people who own Jane's "middle" book when Jane's library is sorted by total members, in this case book #2:
10 median book obscurity
Jane's median/mean book obscurity is 10/100 (again the slash is just a separator, not a division operator).
The difference between the median and the mean obscurity is a measure of how "skewed" the book distribution is. In this case, you might say Jane's library is "more obscure", because her median book is only owned by 10 people; but on the other hand the absolute numbers also mean something; Jane's most popular book is owned by almost three times as many people as John's most popular book (289 versus 100).
In your case, your mean obscurity is 1970:
(sum of all the LibraryThing members who own each book in Oliviine's library) / 356 = 1970
and your median is 31:
members who own book #178 (356/2) when Oliviine's library is sorted by "total members" = 31