Some quick updates:

• The site has been slow, and today there were some server glitches. Can you hear the hamster wheel squeeking?
• A new, much better server is coming tomorrow, or possibly Tuesday. I think it will be a significant improvement. Whether it is or not, the new setup is much more flexible and expandible.
• I spent the day working out how to disambiguate editions. They’ll be big changes in that area soon.

I just want to put something up to express profound disapointment. Google, a company I’ve believed in for years, has decided to give into censorship in exchange for better access to the Chinese market. Google will get a .cn domain, and searches in China will now exclude anything the Chinese government wants.

Google hasn’t gone as far as Yahoo yet—forking over user information that lands dissidents in jail—but with today’s action I wouldn’t put it past them.

For shame, guys. For shame. I thought better of you.

Links: CNN, Boston Globe/AP, Google Google group.

PS: To think, people were congratulating them for refusing to give anonymized user searches to the US government, as if that were a serious issue. (“I support freedom!” one soon-to-be-duped fan wrote on their support group.) Censoring search results in exchange for a little money? That’s a real issue.

As some of you have noticed, I’ve added a “Fun statistics” link on every profile page. Here’s mine. It presents statistics like:

• Library obscurity (the average number of copies of the books held by others)
• Total tags; tagged books; tags per book
• Cataloging sources
• A histogram of ratings
• A histogram of publication dates

There’s a lot more I’d like to do. Right now these numbers aren’t contextualized. Is my “library obscurity,” 12, high or low (it’s quite low). A percentile, and maybe something on the Zeitgeist page would be a good idea.

I’m very open to other ideas. I can add read-date statistics, for those of you using those fields. Ditto BCID fields. I could do tag-obscurity, for what that’s worth. I’d like to work with page counts; I have the data for some books and can get it for others. In theory I should be able to extract publishers from ISBNs (I can do it from the publishers field too, but that’s too messy). But extracting publishers from ISBNs requires a big database, and I think it’s one I need to pay for, from Bowker I think.

Some other ideas would require a more serious integration with library data—language, date of original publication, etc. Nicole on the LibraryThing discussion group suggested “some kind of histogram of the distribution of your books on the LoC classification
system.” That’s a good idea too.

LibraryThing has a handful of competitors. I’m not too worried about them. LibraryThing wins on features, as many blog comparisons have shown (eg., the Powells blog). And LibraryThing certainly wins on size and traffic. The most frequent comparandum started the same week as LibraryThing but has 4% of the books. Another has 20% as many, but it’s been going for more than four years; it hasn’t broken the web’s top 100,000 sites in six months.

I’ve written to most of LibraryThing’s competitors, seeking a deal that would combine our sites, giving all their users lifetime accounts and bringing the developers on board as partners. So far, no dice. I guess it’s more fun to do your own thing.

A few days ago I came across a new site, Bokhyllan.com, a sort of Swedish LibraryThing. Between guessing-in-context (“LOGGA IN”), bad German and a second-degree connection with someone who knows Swedish, I managed to get the jist of it.

What a perfect situation! They’re just starting, and the site doesn’t have all the features it needs. LibraryThing has those features—even integration with the Swedish library system! But I have no desire to develop a Swedish LibraryThing, and no ability even if it did. It seems silly to do all this development multiple times. Why not combine forces and write the code once? Take the code, keep the domain, charge for it or make it free—I don’t care. But share your improvements and we all win.

I wrote. No answer so far. Maybe this plug will get their attention.

If anyone knows of any other non-English LibraryThings, let me know. If you want to start one, let me know. More generally, if you want to help LibraryThing grow, let me know.

There’s a more general conversation here about open-sourcing LibraryThing. I guess I’m a sceptic that it could work. At a minimum, a whole separate site would need to be set up, with fake data. Security would also be a big issue. If it didn’t work, it would be a big waste of energy. If it caught on, I feel like my job would become cat-herding programmers. In the worst case, LibraryThing would end up fragmented into a half-dozen forks.

Maybe we could discuss open sourcing on the LibraryThing Google Groups group. The group is growing in greatness—must keep using words starting with g. If possible, I’d like to keep blog comments on the blog.

All requests eventually come true. Searching has been improved, replacing the old “relevance” search with a more Google-ish search.

The search page now look like:

The main improvements are:

• Searching for “Greek history” now gives you all the books that have the phrase “greek history,” not some lame weighing of books that have just one or maybe just have a word close to “greek.”
• Book searches can use syntax like +greek -history
• You can now search LibraryThing for authors and tags. Instead of dropping you into a catalog with X-thousand books by J. K. Rowling or tagged “science fiction,” these take will take you to the LibraryThing page for J. K. Rowling and science fiction.

Still to be improved:

• Book searches are still ordered by relevance, which means they can’t be sorted by author, title, etc. I’ll fix this soon.
• You cannot currently search the entire 1.5 million books for a single book. I’m going to add this, but add it in a way that you get the book-page, not a catalog page loaded 10,000 copies of Harry Potter. The main problem right now is speed; I don’t want to introduce any more speed-hogging features.

Of course, you can search individual user’s libraries. To go their catalog and click the “search” link. But this doesn’t seem enough to me. Perhaps profiles should include a search box, or the search tab should have some way to select a user to search. I’ll tell you that I don’t want a menu with 21,000 people in it! Ideas taken.

Update: LibraryThing is still mostly up, but sometimes slow. I am working on the server issue on a number of levels and hope to show some progress soon. This is one issue I will lick, no matter how large it gets. That said, if you’re David Pogue and writing up an article on LibraryThing for the New York Times, please don’t publish it this week!