Wednesday, May 30, 2007

1pm Abby talking at BEA / NYC Meetup

If you're at BEA this week—and what book-industry type is not?*—come check out Abby, LibraryThing's first hire and Head Librarian give a short talk on Thursday at 1pm.

Also, she's organizing a get-together in NY. Friday night at 6:30, anyone and everyone—BEAers or not—is invited to meet up at The Half King Bar & Restaurant (505 W 23rd Street).**

Abby's speaking alongside representatives from HarperCollins, Grand Central Publishing, MySpace and Gather. The topic is "Using Social Networking to Build Author Brands."

She's going to outline what LibraryThing is all about, and how authors and publishers are using it. But LibraryThing is something different—more? less?—than "social networking" and "author brands" is one of those bloodless, push-push, container-shipping phrases that obscure what's really going on. Readers don't connect with "author brands" anymore than passionate lovers connect with "lover brands." Social networking--or social cataloging--is about real connections. Brands are to real connections what television is to telephone.

Anyway, quibble aside, I'm sure it'll be a great panel.

*That would be me. I'm on a 1x3-mile island off of Ireland. Really.
**The Half King is apparently a very literary place—they have readings every Monday night, and it's co-owned by Sebastian Junger. If his place is full of LibraryThing-ers, surely he'll become a LibraryThing author.

Monday, May 28, 2007

Bad news from Bookland

Two depressing stories:

Saturday, May 26, 2007

Excellent: David Weinberger at Google

David Weinberger went to Google and did his Everything is Miscellaneous talk. It's now posted. Unlike all the other talks I've seen—David has a dozen or so here and there on the web—they preserved the David's inimitable, often hysterical, "slides."

Recommended—no essential—watching for librarians, information architects and so forth.

Friday, May 25, 2007

LibraryThing bests Vanatu

I get a lot of LinkedIn requests. Today's came with a factoid:
Fact: More people have joined LinkedIn than live in Sweden.

According to this website, that means LinkedIn has more than 9,001,800 members. Which leads me to the following LibraryThing factoids:
Fact: More people have joined LibraryThing than live in Vanuatu.

Vanatu, baby, Vanatu. Not some teeny-weeny place like Tonga, Andorra or Liechtenstein. Vanatu!

Wednesday, May 23, 2007

Hippo Birdie Two Ewes, Linnaeus

It's his 300th birthday.

Saturday, May 19, 2007

David Weinberger at Yahoo

I've got some unfinished David Weinberger business, including a review and a pile of free copies of Everything is Miscellaneous to give out. (I'm thinking some kind of contest, like photos of sock drawers, miscellaneous piles, or interesting ways to sort things.) Until I have a spare minute for that, here's a great video he just did at Yahoo.



PS: I would like to lodge an official complaint against Yahoo. I was watching it on Yahoo Video itself, and half-way through. The page invited me to rate the video and I clicked it. Well, that was dumb! It ended the page and called up a new one, asking me to sign in. Well, that's the last time I rate a video on Yahoo!

Freeze—I'm a change agent

I had the honor of key-noting at the Southern Maine Library District's Spring Council meeting. Until this my contact with local librarians had been pretty meagre, so it was great to meet so many, and be received so warmly.

I'm particularly keen to do something with one of the librarians I met from Maine's island communities. It would be great to get the small island library in, and then all the private books, so residents could search for them. Apparently island librarians are already doing some of this informally, in their heads.

I've been interested in "small scale" social software ever since I read Clay Shirky's Situated Software. Small-scale social software can dodge some of the problems of large-scale varieties. So, for example, sharing books between LibraryThing members would require a complex "reputation" system to deal with bad apples. On an island with 200 people you don't need that.

For keynoting I received a Maine Libraries mug—nice, but too slight for my super-sized coffee addiction—some chocolates—inhaled!—and the pin to the right, "Change Agent" on a sheriff's badget.

No doubt it was either that or a round button, and the sheriff's badge was cooler. But there's something richly ironic in the pairing of "change agent" and a sheriff's badge. Does change come from the sheriffs? Maybe it's a deputy badge, or whatever you give someone in a posse.

I keep hearing variations on: "I hate my job. I have no power. But I'm going to stick it out. My supervisor is going to retire some day."

Man, that's gotta suck. Startups have their downs too, but that's a kind of pain I hope to never feel again. Maybe the library world needs more villains and vigilantes. Damn the law, and string up that OPAC!

Friday, May 18, 2007

Will Vertical Social Networks Succeed, or, Is Goutster a big deal?

This is a response to Big or Bullshit: Vertical Social Networks, posted by David Cohen's to his blog, Colorado Startups. Cohen examines vertical social networking, and concludes:
"I think that the vertical social networks will proliferate and become a useful part of the web. But in terms of their potential as an investment - I guess I’ll have to call Bullshit. Look to the aggregators and the toolsets that emerge around identity - I think that will be Big."
As founder of LibraryThing, I always chafe at the idea that books are a "vertical." I mean, books are the world. Connecting around books isn't some marginal activity—it's practically civilization itself!* But I guess that marks me as a vertical-dweller. Maybe the running people think that books are a luxury, and running the center of life—the Philistines! At least you didn't compare us to DwarfDate or that network for people who really enjoy sneakers.

I think we'll find some verticals are a lot more promising than others. A social network for sufferers of gout or mesothelioma—to take the most expensive Adwords term—could be quite profitable.** The network for sufferers of Guinea Worm will not. The situation may end up resembling specialized periodicals. Mainstream magazines go up and down; it's vicious market. But there are a lot of tiny publications we've never heard of that chug along year after year making a killing in smaller, profitable spaces.

But I don't think we know all the ways social networks can make money. There are surprises in store. You're assuming the center is advertising.*** But LibraryThing, for example, HAS no advertising. Not only do we offer a service good enough to charge top users, but the data members add--distilled statistically into recommendations and so forth--has a lot of value. A better recommendation algorithm is a force-multiplier for ecommerce. And just a few days we announced our first library license, putting LibraryThing in the online catalog. Libraries aren't rich, but they spend a lot on technology. And there are more libraries in the United States than McDonalds. I don't know how this applies to other verticals, but I expect advertising is just the beginning. (And for the sake of our exhausted eyeballs, I hope the ads go away.)

I can't get behind your enthusiasm for social network aggregators. The verticals are about passion. The Buzz page on LibraryThing is a florilegium of love letters. As a book guy, I understand this. I'd feel the same way if LibraryThing weren't my baby. But I've never gotten excited about belonging to a social-networking aggregator. I don't know anyone who has. There's probably a business there, but a low-margin one, like providing free RSS aggregators. And, by definition, you can't have as much lock-in when you're not the thing they really want.

I wonder if the interest in aggregators isn't just a repeat of a familiar pattern, when the internet discovers something outstanding and fun, and VCs later decide that, although fun doesn't pay, a boring version will rake in the dough. Remember when "business-to-business" ecommerce was going to save the dot-com bubble?

Even if verticals like LibraryThing's aren't going to make for "big exits" and--as I also believe—most verticals don't need VC-level funding****, there's something going on here that deserves VC attention. Vertical social networks aren't some fad, but the virtual life of something even deeper than books—the need to connect with like-minded people. If, as Warren Buffet has said, the airplane industry--taken as a whole over the last 100 years--hasn't made a profit, it still changed the world. A changed world is a world with opportunities.

*Exempting Last.fm is particularly galling. The book industry is four times larger than the music industry! Okay, I made that statistic up. It's probably the reverse. We need to hire a marketing person who can look that stuff up.
**The gout network would carry a lot of sherry and madeira ads, of course.
***And don't get me started on affiliate revenue.
****There's also a neat paradox with "exits." The most successful verticals will be, like LibraryThing, created by people more interested in the vertical than getting "out" of it. That's one reason LibraryThing didn't take VC money, so I'd be the one who went for an exit deal, not you.

Labels: , , ,

Why I joined OCLC ...

... is the title of a short Library Journal piece by Roy Tennant. In it, Roy, a popular and much respected library speaker and author explains his decision to leave the California Digital Library and take a job at OCLC.

Roy's decision drew some flack among anti-OCLC librarians and related pundits who view OCLC as—in Steve Oberg's phrase—the "Microsoft of the library world."

I'm in that camp, as Roy knows well. After we presented in the same session at Computers in Libraries Roy and I went out to dinner, with another prominent librarian. I subjected them both to a long, Greek-food-fueled rant about open data and the problems with OCLC and its approach to the web. I had my shot. A day or two later, he announced he was moving to OCLC. Apparently I didn't convince him! :)

OCLC needs people like Roy—passionate librarians with a vision for the future. If OCLC is to change, people like Roy are going to be the ones to do it. I have great hopes for him there.

But they're not going to do it alone. People on the outside are going to change OCLC too. They're going to keep the pressure on. I applaud that Roy took the time to explain his move, and his vision for OCLC and the future of libraries. He was eloquent and persuasive. But I'm also glad he felt he had to.

Labels: ,

Tuesday, May 15, 2007

LibraryThing for Libraries gets MUCH faster

The Tag Browser in FAAAAST!
The top complaint about the LibraryThing for Libraries was the speed of the tag browser. We revisited it today, and made it MUCH faster. It's positively zippy!

It proved to me, yet again, that sometimes your seemingly perfect MySQL query—JOINS across five tables on primary keys and EXPLAINs like a dream—can be broken into two queries for an order-of-magnitude speed increase. If I were smarter, I'd understand this.

Monday, May 14, 2007

Danbury, CT kicks off LibraryThing for Libraries!

Danbury's catalog, with LibraryThing for Libraries in place.
The Tag Browser in action.
We're happy to announce that the Danbury Library in Danbury, Connecticut has become the first library in the world to put LibraryThing for Libraries on its live catalog. The Danbury Library—already breaking ground with an active blog and a MySpace page—continues to innovate and experiment. And we finally have something to show people!

We hope the library world will give it a look. It's not the full and final "next generation" solution. It's not perfect. But it's something new, and--we think--quite promising. It's also available now, and only going to get better.

Underneath, the data comes from LibraryThing and its members. We've cleaned some of it up--Abby, our head librarian, and Jenny Anastasoff, a local librarian who's intern with us, have been hunting down and excluding personal or irrelevant tags.* But its strength is the strength of LibraryThing's people and their collections—200,000 members, 13 million books and 17 million tags.

An example. Let's start with Helen Fielding's Bridget Jones's Diary (http://cat.danburylibrary.org/search/i?SEARCH=0670880728), the well-known comic novel which kicked off the "chick lit" craze. The original record provides little more than the author and a single--rather useless--subject "Single women--England--Fiction."

LibraryThing adds three pieces of data:
Other editions and translations. This picks out other editions and translations of the item. In this way, a patron who discovers that the item they searched for is loaned out can discover another, available copy. The feature also creates reciprocal links between paper and audiobook editions, and between different language editions. Underneath, LibraryThing is drawing on its "works" system, a simplified version of FRBR.

Notably, we're making this feature free. You don't need to buy the other widgets and, at least for now, there's no setup cost.** There has been a lot of buzz lately about "FRBR-izing your catalog," and I've seen some very high prices quoted. OCLC's xISBN API just graduated from free to paid. We're keeping it free, and serving up features, no an XML file.

Similar books. This too is based on LibraryThing's data--the holding patterns of some 13 million books. It is not, of course, perfect. It makes mistakes--even a few howlers. But, overall, it's as good as Amazon's recommender system, and particularly well-suited to libraries. For starters, it tends to move patrons "down the long tail" or, to use a more appropriate metaphor, "into the stacks." Thus, C. S. Lewis' The Last Battle offers up two other Narnia books, but also Madeline L'Engle and Susan Cooper. Shelley's Frankenstein links to Dracula, but also to The monk by the Victorian Gothic novelist Matthew Gregory Lewis.

The Cyberpunk tag.
Tag Browser. This calls for a longer explanation. In essence, the Tag Browser allow patrons to examine Danbury's catalog using simple, common-sense terms. The terms, or tags, are drawn from LibraryThing 17 million member-added tags.

Bridget Jones's Diary
is a good place to start. Everyone knows what it is--it's chick lit. But there is no "chick lit" subject in Danbury's catalog--or in anyone else's. Even so, chick lit is "real"; patrons will want to look for it.

Knowing that a book is tagged "chick lit" isn't necessarily that helpful. But click on the tag, and it opens the Tag Browser for chick lit. The list it gives amounts to something like "The Chick Lit Reading List"--from Lauren Weisberger and Jennifer Weiner to Sophie Kinsella to Emma McLaughlin, all available at the Danbury library.

The same works for many other terms. In Danbury, the seminal cyberpunk novel Neuromancer gets no subject at all. In the Tag Browser, it shows up at the top of "Cyberpunk" (which is another good "reading list"). Even the "Science Fiction" subject falls short. Only 291 items get the subject***, which don't include Dune and Martian Chronicles (which have no subjects) or Heinlein's The Moon is a Harsh Mistress gets by with "Imaginary wars and battles--Fiction."

Tags are not always better. Subjects' authority and potential for hierarchical presentation cannot be dismissed. And, unlike some other systems, Danbury very effectively aliases common terms, mitigating some of the worst problems. Thus, "Chinese Cooking" links to a search for "Cookery--Chinese," "Ice age" maps to "Glacial epoch" and "Prozac" to "Fluoxetine." The latter picks up some items the Tag Browser tag "Prozac" does not.****
But it goes the other way too.
  • The subject "PHP (Computer program language)" misses Dreamweaver MX : PHP web development, classed soley under "Dreamweaver (Computer file)."
  • The "MySQL (Electronic resource)" subject misses the High performance MySQL—far and away the best MySQL book in the building—but under "SQL (Computer program language)."
The Usability tag.
Also with the Tag Browser, you can examine Danbury's excellent holdings for the hot and quite distinct topics "Usability" and "Information architecture." The former includes 13 books, the latter eight. Without tags, they are just part of the much larger "Web sites--Design" pile.

Or take "Mesopotamia." A Danbury subject search turns up two books of juvenile fiction, and no cross references. To get the epic of Gilgamesh, you need to known the trickier "Epic poetry, Assyro-Babylonian--Translations into English." Other key works, like "The Cultural atlas of Mesopotamia and the ancient Near East" and "Cultural atlas of Mesopotamia and the ancient Near East" fall under Iraq subjects. This isn't a bad place for them, but multiple access points are always good.

In short, by adding LibraryThing for Libraries, Danbury has opened up some valuable new ways for patrons to find books of interest to them.

How it works. Perhaps the most remarkable thing here is the relative ease with which Danbury added the feature. Danbury has an Innovative Interfaces OPAC ("WebPac"), but LibraryThing for Libraries is platform agnositic. There is no back-end integration--no one at Danbury had to call their vendor. Ultimately, it involved just a few lines of JavaScript pasted into the page footer. It works on any modern browser. It causes your tea service to fill with magical gold. Okay, not that.

Caveats.

Some things we're working on:
  • So far, it only knows ISBN books.
  • The Tag Browser is not fast enough. This can be improved, no question.
  • There are some wrinkles in the tag data, particularly in marginal cases. Some of this will go away, but not all of it.
  • There are some character problems, like the tag "C++" turning into "C."
  • In the rush to get the feature out, we forgot to give Danbury the "noscript" tag, so that users with JavaScript turned off can enjoy the enhanced content.
Improvements:
  • We wanted to start with non-interactive features. We love user tagging, but wanted to show what out existing data could do without any additional patron data. There are also sticky privacy issues there. But, of course, this is a direction we'll be moving in.
The Future. Getting one library out takes a lot of pressure off us. We've been doing a lot of custom work to show potential customers what we're going to offer. With a major test case live, we can cut back on that. We've also learned a lot more about the problems of importing library data, and are making progress there. We'll moving into "production" mode, and getting back to other libraries soon.

Interested? If your interested in seeing LibraryThing for Libraries in your catalog, send us an email at tim@librarything.com or abby@librarything.com.

Thanks. Altay, John and I want to express our thanks to Kate Sheehan, Danbury's Coordinator of Library Automation, who pushed Danbury ahead of the LTFL pack through sheer enthusiasm and force of will. (Also, she checks her email constantly, which fits our work habits nicely.) Kate has an entertaining blog, Loose Cannon Librarian.

*For example, the most of "personal tags," like "to be read" and "at the beach house," have been removed. This is only about what shows up in the library catalog, obviously; nobody has been tampering with user data on LibraryThing!
**We've reserved the right to start charging a setup fee, if we get a huge rush of non-payers. So long as we can do it easily, we do not want to charge anyone. Either way, LibraryThing's underling isbn-to-isbn data will remain free and open for all non-commercial use. This is in stark contrast to OCLC's parallel xISBN service, now restriced to 500 queries a day.
***Another 147 get "Science Fiction Gsafd," surely one of the best acronyms ever developed--"Guidelines on Subject Access to Individual Works of Fiction, Drama, Etc." Unfortunately, the two subjects don't sort close together.
****Not, however Elizabeth Wurtzel's famous Prozac nation : young and depressed in America. Wurtzel's book gets two subjects, "Depressed persons -- United States -- Biography" and the wonderful "Wurtzel, Elizabeth -- Mental health." Hey, how do I get a top-level category, and another for my problems?

Monday, May 07, 2007

"Why Web 2.0 is leading back to full cataloging"

Interesting post on LibraryJuice on "Why Web 2.0 is leading back to full cataloging" covering LibraryThingThing and Last.fm.

The quick summary--okay, the first sentence and the last!--is:
"We often think of Web 2.0 sites in terms of the idea of “tagging instead of cataloging.” ... Free-form tagging has its place, but where consistency and accuracy counts, as it does in many Web 2.0 sites, I think reliance on users will turn out to have been a dead-end, and there will be a new appreciation for our professionalism."
I agree with the idea that Web 2.0 can—indeed, in LibraryThing, has!—lead to a new appreciation of library data, libraries, librarians and catalogers. It might help, however, if more libraries (and particularly OCLC) released their data, and in friendly formats, not MARC. (It's hard to appreciate something you don't know about and can't use.) But I strongly disagree that there's a real tension between the two—that tags are the enemy of subjects, for example. But it's certainly food for thought.

If the blogger reads this, how about spending some more time on LibraryThing? Some of LibraryThing's social cataloging features were mentioned, but then not brought into the argument. For example, other librarians have noticed that LibraryThing's work-disambiguation data was of very high quality, and a number of libraries are already using it. That data is member-created, but built on library data. Maybe that's the future--a constructive collaboration and mutual respect between professionals and amateurs.

Sunday, May 06, 2007

Funny cause it's true?

"Somewhat Obsessive-Compulsive"—found on a MetaTalk thread about LibraryThing.



(Posted with permission of John C. Ralston, of A Year in Comics)

Thursday, May 03, 2007

Combined blog feed available

I used Yahoo Pipes to make a combined feed for this blog and our Thingology blog. It was easy to do, and the result is pretty useful. The three feeds are as follows:
Unfortunately, if you're already using it, you got this message twice. From now on, I'll be less likely to cross-post.

I also edited the employee list on the right, to add Altay. He is the magic behind the LibraryThing for Libraries Javascript, but almost nobody's seen that yet, so we're waiting for his first user feature to give him a proper introduction.

Wednesday, May 02, 2007

Take my wife, please

The Vermont College literary magazine Hunger Mountain is eBayin-off writing critiques by its professors and alumni. These include Bret Lott (wrote Jewel, an Oprah book), Poet Maxine Kumin and prize-winning and exceptionally talented novelist Lisa Carey, my wife. All proceeds go to support the magazine.

We should get find some librarians, booksellers and subject specialists to auction off LibraryThing library critiques...

LibraryThing Tag Consortium

Since Abby mentioned our plan to for "tag consortium"—think OCLC, but for tag data, and without licensing restrictions, I get to show the logo I put up at Computers in Libraries:



Spent half the night trying to get the fonts right, instead of writing my talk...

PS: Abby relates that she saw an actual kookaburra, sitting—you guessed it—in a gum tree.

Superpatron advice list

Ed "SuperPatron" Vielmetti has posted an excellent list of "Ten ways for superpatrons to help build better libraries," for an upcoming talk. Blogs love lists, and this is a good one. (LibraryThing is number five.)

Tuesday, May 01, 2007

Kangaroos! (and the NLA)

I'm finally back from Australia and a whirlwind week of vacation in California, and I'm rearin' to go!* I had a fantastic time speaking at the National Library of Australia's Innovative Ideas Forum. It was a lively mix of speakers, including Susan Chun who talked about the steve.museum project—tagging art! Very cool. Canberra is a beautiful city, and the NLA bent over backwards with hospitality (the kangaroo picture I took in the backyard of NLA's head of IT!). I'd go back in a heartbeat :)

My talk focused on tagging and the wiki-like aspects of LibraryThing.** It was an incredibly receptive group—I'd never seen so many librarians clamouring for tags! There was a serious push to get the "tag consortium" that we've talked about before up and running. Because, as Tim has mentioned many times before, tags are most useful when they’re available in large quantities. So we're going to offer up APIs and widgets that will allow libraries to both add tagging to their site, and to access LibraryThing's 16 million.***

The white building in the background is the National Library of Australia. I couldn't resist the rainbow picture!


update: I forgot! I promised Fiona (who called me "the biggest geek there") that I'd bring attention to the group she created right after the program, Aussie Librarians. So go join!

Now, to get through all the email piled up in my inbox...

*Advice for life: having a vacation back-to-back with a business trip is exhausting. Particularly when time-travel is involved—Tuesday the 17th never happened for me, but Saturday the 21st I lived through twice. One 4/21/07 I spent walking around Sydney harbor (thanks to a 10 hour flight delay), and then I repeated the day on the plane... I have a newfound respect for that Groundhog Day movie...
**Stop and think about how much of LibraryThing is done by its members (i.e., you). You add books, of course, but also tags, reviews, and ratings. You upload cover images of books, photos and pictures of authors. You combine tags, author names, and works. You add links and disambiguation notices to author pages. And you've translated the site into 30 different languagues. That, folks, is nothing short of incredible. The helpers page on the Zeitgeist, if you haven't seen it, is an addictive chronicle of all this work.
***LibraryThing for Libraries is certainly a big step in adding tagging to library catalogs (we've got tag-based browsing widgets up and testing now, and are are working on widgets that will allow patrons to add tags).