I’m not making major changes this week, but just planning them. (I’m traveling on business.) But I thought I’d respond to some comments.

OCLC. I’ll look into “find in a library” functionality. I see the attraction.

Tags. I don’t think I’ll split different sorts of tags. Tags should be absurdly easy to enter.

Peter. Shoot me your email again. I try to respond to everything, but some emails slip by. I am interested in your UI ideas.

Covers. Covers are determined by ISBN. What LibraryThing should do—I think—is list all the ISBNs and covers for a given “book.” I’ll add that when I add edition disambiguation.

Christian Science. Google it. It’s a church—The Church of Christ, Scientist—started in Boston by Mary Baker Eddy. Among other things (and at the risk of speaking for another’s religion) Eddy believed in faith healing—that illness was really spiritual. The newspaper is connected to the church, but the only concrete result from that is a daily column on Christian Science themes.

Thanks for all the positive comments. I’ll be back making daily changes soon.

If you’re wondering why LibraryThing is a little slow, it’s because traffic has surged. LibraryThing was picked as the Kool Site on the Kim Komando radio show. I regret to say that I’d never heard of Ms. Komando or the show, but apparently she’s huge.

Anyway, more than 1,000 users signed up in the past three hours (a 10% increase!). Most are just checking it out, but the traffic is still pretty staggering. I suspect it will pass in a few hours, leaving a few hundred dedicated new-comers. To them: Welcome!

Bibliophile web geek Peter Harkins (LibraryThing: Malaprop) has created a sweet Firefox extension called SelectThing. SelectThing allows you to search LibraryThing from any web page. Peter writes:

“I’m a big fan of cataloging books I want to read on LibraryThing and often run across mentions of interesting books while browing the web. But it always distracted me to copy the book’s name, open a new tab, pull up LT, go to the add page, paste in the text, hit search… So, like a good geek, I wrote a Firefox extension to simplify it.”

Here’s how it goes: You’re reading a book review on the New York Times or your favorite blog and you want to add the book to your LibraryThing catalog (soon wish list too). You just select some text—maybe the title of the book, the author or an ISBN—right click and choose “Search on LibraryThing.” It opens a new tab with LibraryThing’s “Add Books” screen, and does the search too. It’s super quick and easy.

The extension is online at http://selectthing.mozdev.org (see more screenshots). The install process is standard (and easy). There’s a mailing list there, but Peter will also be monitoring this blog for feedback.

Let me be the first to offer mine: what a cool idea, and nicely done!

Some recent fixes and improvements:

• I finally fixed the author cloud, both individual and the all-users cloud.
• Authors (eg., J.R.R. Tolkien) now include a “commonly tagged” section.
• Recommendations have been improved, incorporating a progressively severe bias against books by the same author. Recommendations remain randomly excellent and occasionally crazy (see *).
• Author links now work; some were “redirecting” in circles.

Own, read, want?

UPDATE: People are clearly confused by my categories. For the record, they were “(1) books you own, (2) books you’ve read but don’t own, (3) books you want.” I condensed that verbally to “own,” “read,” “want.” “Read” is causing people to think I’m trying to capture to distinguish how many books you’ve read (including in your catalog). I’m trying to allow people to feel comfortable cataloging books that aren’t currently in their library. Clearly this terminology isn’t going to work. Any suggestions?

I’m working on adding a “collections” feature (name in flux and up for suggestions). At present LibraryThing presumes you own your books, although users have added tags for books they’ve read but don’t own and for books they want. With the holidays approaching, I’d like to add the latter, at least. But I haven’t decided on the approach. Here is my plan. Executing it require bringing lots of features online at once, not making incremental changes and seeing what you people think as I add each one. So here’s a sneak peak. Comments would be appreciated.

I want to have a small number of fixed, non-overlapping categories. If users can define their own categories people will use “collections” instead of tags—for “at the beach house,” “read but hated,” etc. This will satisfy power users, but reduce universal value. Keeping the collection “buckets” limited will make it easier for people to view and understand other’s collections. For example, the default view will include everything a user owns or has read, but profiles will have a link to users’ “wish list.” Predefined terms are also better for library-wide statistics—whose wishes are like mine, what users are wishing for generally, etc. This won’t work if some people call it “wish list,” others “want” or “christmas,” etc. I don’t want to get into users picking their own name and then giving the collection a “type.” The system must be transparently simple.

I propose three collections: (1) books you own, (2) books you’ve read but don’t own, (3) books you want. I’d like to use the short names “own,” “read,” “want.” But people may misunderstand “read.” “Want” is my attempt to avoid the Amazon-y “wish list.” “Own,” “read,” “want” is also gramatically parallel. But can I escape this well-known name? Jacob Nielsen talks about a company replacing the standard “shopping cart” with a “shopping sled”—purchases plummeted!

I plan to implement it so that you choose what collection you’re adding to on the “add books” page. (I’ll have to think about how it works when you add a book another way; I don’t want to make things too complex.) There will be an easy way to “power edit” books into one collection or another. The catalog view will show “own” and “read” together by default, but allow you to choose to see any combination of the three. Profiles will break it down three ways.

That’s my plan. As they say, no battle plan survives contact with the enemy. In this case, the enemy is the PHP programming language, not you, so help me out!

* I was shocked to discover that the 1969 parody of Tolkien, Bored of the Rings, has as its top recommendation Marcel Proust‘s In Search of Lost Time. Laugh all you want, but eight of thirty-two copies of In Search of Lost Time are owned by people who also own (the obscure) Bored of the Rings. You cannot argue with statistics! But why? I would also like to note that Amazon doesn’t present a single suggestion for Bored of the Rings, and that most of LibraryThing’s suggestions—Proust aside—are fantasy novels. LibraryThing beats Amazon again!

UPDATE: See the author page for National Lampoon. Bored of the Rings is split between three editions that LibraryThing isn’t “combining.” (User-contributed combination is on the way, I promise!) The main one has a review by wenestvedt, and picks out one really good suggestion—William Goldman’s The Princess Bride, a fantasy/Medieval romp rather than a parody.

Various places (Slashdot, Techcrunch, News.com) have reported that Amazon is now experimenting with tagging. Only some customers see the option, and I’m not among the chosen ones. I’m dying to know how Amazon tags are comparing against LibraryThing ones!

I’m not worring too much about this. I don’t think people choose to use LibraryThing just to tag books. To the extent that tagging is the attraction, it’s about tagging your own books. In less than three months LibraryThing users have applied more than 1,150,000 tags to their books. People will do amazing things with their own collection.

Will Amazon get the same kind of buy in? I’m a big-time Amazon customer, but tagging books on Amazon seems to me like volunteering to fluff pillows at the local Sheraton. I suppose if someone has an enormous number of items on their wishlist they will want some memory aid—the most important thing about tags. Absent that, I just don’t see what Amazon customers will get out of it. On the other hand, Amazon has so much traffic that maybe the altruistic, leisured 1% will quickly fill Amazon up with tags.

Another worry. Until now, tags have not had much commercial value. I doubt that the Paris Tourism Board is spamming 43Places. But Amazon has seen review abuse before—not a few authors are livid at the practice. Will Amazon tagging lead to the introduction of the spag (spam tag*)?

Of course, at LibraryThing, despite enormous financial incentive to promote my wife’s novels, I have yet to engage in any unfair tagging of her absolutely terrific works, The Mermaids Singing, In the Country of the Young (described by the New York Times as “exquisitely doomy“) or the recently-released Love in the Asylum, for which I designed the hardback cover. No, we will never do that here.

*Go, my pretty new word. Go go! Settle the world with your beauty.

Here are some responses to the responses to the last post, with some dreaming at the end.

1. Fixes. I fixed the author numbering problem, the author link problem, the apostrophe problem, and some others. There can still be a difference between copy numbers because the author page doesn’t count duplicates.

2. New feature. I added an author rating feature, based on book ratings.

3. Secondary authors. Yes, it’s too bad that secondary authors are not in the current system. It’s a little wiggly, starting with the programming standpoint. There’s also some issues between “second authors” and “included authors.” I’m going to let this run for a while before thinking about it. More generally, I want LibraryThing to handle “contents” in a rich way—to allow it to see the stories within the story collections. Authors are part of that, but not the largest part.

4. Non-suggested authors. Yes, I’m going to add something for the Ratzinger/Benedict or Clement/Twain problem. I think I’m going to restrict it to paid users, in case someone decides to run through the system combining popes with humorists. I’d also like the UI to use an auto-complete, which will take a little work.

5. Author cloud. This was the necessary first step to a fixed author cloud. The author cloud was shut down for a reason—lots of complaints. Chief among these was that the cloud included “the same author twice.” Opening up author names in this way make it possible to fix that.

6. Book-combination. Clearly this is a trial-run for other user-driven features. There is a method to my madness!

In case you read this far, I will tip my hand a little—and dream a little. Some day soon I’ll post more on this topic, but here’s some initial thoughts.

In general, the more open the system the better it will satisfy users. At the same time openness can create real value. That irritated person who clicks a button to make their German and English Harry Potters show up under the same author is forging a piece of information that Amazon doesn’t know. (Actually, they may know THAT one, but you see my point.)

The potential is huge: LibraryThing could be for library catalogs what Wikipedia is for encyclopedias. That’s a little imprecise. By and large, LibraryThing users don’t enter their own data, but take it from Amazon and libraries. But the connections between the data—author aliases, book aliases, contents, subjects/tags, etc.—can all be user-driven. Even now, recommendations, related tags and so forth are user driven—they’re driven by user statistics.

I believe that over time LibraryThing users will generate significant cataloging value. Some of the value will be user-generated. Some of it will be statistical. Most of it will be probabilistic rather than binary. And in the end, LibraryThing will never be the best catalog in the world. But I think it can produce data and will make even THAT catalog better.

All that from the “competitive bastard child of bibliomaniacs and pro wrestlers“!

Sharp-eyed users will notice a new feature—author pages. Author pages, linked to from within your catalog and elsewhere, list an author’s books, who owns them and so forth. This is a toe-in-the-water announcement. Coming days will see more features, including “similar authors” and a Wiki-like ability to add author home pages. Suggestions are, as always, encouraged.

There’s something daring in how the feature is implemented: users control the catalog. Some of you have noticed problems with two editions of the same book not always being counted as the same. Well, the problem is even worse with authors. Not only does LibraryThing draw on five Amazons and 30+ libraries around the world, but as anyone familiar with Library of Congress searching will tell you, author names often vary within a single catalog. Is Joanne K. Rowling the author of Harry Potter und der Orden des Phönix the same as J. K. Rowling the author of Harry Potter and the Order of the Phoenix? A computer can’t tell, but I’ll bet you can.

In fact, the computer guesses pretty well, presenting a list of likely aliases for in the “Also known as…” section. You can check these authors out by clicking on their names. If they’re really the same, and you’re feeling generous, go ahead and click “combine.” The authors will be smoothed together, with the more common name winning. I’ve gone through some of the better-known authors—the rest are up to you.

Be bold! The system is self-correcting. If you screw up and combine two authors who aren’t really the same—eg., Thomas Wolfe and Tom Wolfe—someone will notice. Clicking “separate” will break them apart again.

Finally, although combined on a global, statistical level, user data is NEVER changed. Call her Jenny Rowling for all I care. Nothing’s going to change that data.

Again, they’ll be some improvements, but the core functionality is there. I’m eager to see how it gets used.

LibraryThing user Sluggo had the great idea of setting up a LibraryThing Frappr map. Frappr is a fun, light-weight tool for putting things on Google maps, in this case LibraryThing users. If people use it, it’ll give a clearer picture of where users are, and give people more reasons to connect with each other.

It’s totally optional—you put yourself there, not the LibraryThing system, nor does LibraryThing control the data. It’s all by zip code, not actual address.

Check it out at http://www.frappr.com/librarythingusers . I propose your “shout out” contain your profile URL, at least.

It took a while, but I finally solved the issue with using ISBNs on non-US Amazon. Sorry it took so long.

I’ve added a user rating section to a book’s “social data.” Here’s the DaVinci Code‘s:

More fun data coming up!

Tomorrow’s edition of The Christian Science Monitor includes the article “Do Your Own LibraryThing” by Jim Regan, already available online. It’s a sunny, detailed look at LibraryThing, “poised to turn the cataloging of books into a form of communal recreation.”

With luck some the article will spawn others, but I’m particularly happy they got to it first. The Monitor‘s a great newspaper—international, in-depth and analytical. It’s been knocked about a bit recently, but it has made great strides on the web. The LibraryThing article shows CM‘s strengths: instead of a glib focus on library-size competition or the seeming dominance of J. K. Rowling (see below), Regan describes the site in detail, like someone who’s actually used it. The reader can make up their own mind.

So, hats off to Regan and welcome to Monitor readers—send me an email and tell me what you think of it!

J. K. Rowling and the popularity myth

I’m going to set it to show more of the top authors for context. People do have a lot of Harry Potter books but people have a lot of books generally—Rowling is only 1/3 of a percent of the total. In fact, LibraryThing exposes the popularity myth. For example, while Dan Brown seems so popular right now, his 379 books are beaten by Umberto Eco’s 476 and trounced by Jane Austen’s 736. C. S. Lewis flings him down and dances on him—1,706!

When’s the last time you read a newspaper article entitled “Dan Brown is selling well this year, but it’s not that big a deal in context”?