LibraryThing has hit 1.5 million books cataloged and 2 million tags, all since August 29.

What used to be a straight line of 9,000/day has become a bending curve—gulp! Even so the site pretty-much stayed up today. The database isn’t out of the woods yet, but it’s in better shape. You’ll see some stutters, but downtime should be minimal and I’ll be getting a second server for some of the “thinking.”

Some other updates:

• Profile pages now show ten random books from the user’s library. People were doing this with widgets, but that presented some security problems and also looked bad with the new design. I will probably add the ability to customize what people see (most recently-read books, and so forth).
• Acquisition, started and finished dates have been added. This feature hasn’t been extended everywhere—eg., you can’t use it in a widget. That will come soon. You can enter dates in North American or European format, but the catalog currently only shows them in the righ—um, North American—way.
• I added a statistic to the Zeigeist page listing the “50 ‘completist’ authors.” These are authors that, if you have one of their books, you have a lot. Think people who collect every single Agatha Christie, although the actual authors may surprise you. To qualify the author needs to be read by at least 100 users. Come to think about it, I could do a statistic for whose library is the most “completist.” I am quite sure that a science fiction fan would win.

It’s up again. Knock on wood, I think it will stay that way.

In general, the system is stronger and more disaster-resistant. I am working with not one but two competent database administrators. There will be some glitches, but also significant improvements. I am confident that there is a clear path to scaling this thing up very large.

There are some changes, some temporary, some permanent. Most of the heavy-duty processing has been moved away from the page-display. So it does not calculate the book suggestions when you enter a page, as it did before—although it then cached them. Now it gives you the cached suggestions if there are any, and no suggestions if there aren’t. The suggestions are being generated behind the scenes when nothing else is going on. Soon enough all books will have recommendations again. Since they don’t do much “thinking” the book pages are much faster.

You’ll also notice that the profile page looks different. If it’s your profile you get a lengthy list of who shares books with you. If another person’s it fronts the list of shared books, something that was pushed down the page before. It doesn’t give you a list of that person’s list of top sharers. I’d plan to add these other-profile statistics back in soon, but they will be subordinated to the shared books list, which is, I think, the most interesting piece of information.

The profile page currently lacks links to user reviews and the more intensive similar-libraries calculation. These will be back as soon as I decide where to put them. I plan to make the profile a multi-page affair, moving the comments off the main page. Those of you who have dozens of comments will be glad of that.

That’s the news. Thank you for all your support. It was so supportive that I plan to make downtime a weekly event.

I’m sorry about the downtime. I have reached something of an impasse. Something is just different, and I haven’t been able to figure it out what, nor my various helpers. Suggestions to increase normalization or change to a different db engine are really beside the point. There is some reason it worked two weeks ago and won’t work now. The number of books is not much higher. No new feature has been added. In fact, they’ve been taken away.

If there is anyone out there who is confident they could help—it’s a PHP, MySQL site—give me a ring. I’m quite willing to pay $65/hour for competent help.

10:06 – Widget requests are now going through.

UPDATE: Okay. It’s up. Let’s see if it stays up. I’ll blog about various changes soon. You’ll notice some missing recommendations. The site will go down at 3am EST.

I’ve got good news to report on the database front. I’ve made something of a break-through. On the nastiest queries—for example, determining what 20 users share the most books with you and how many they share—it now performs 500 times faster than before. You probably won’t see that sort of increase, but it will be noticeable. If you have more than 1,000 books, visiting your profile will become fun again, not an invitation to wait. And the changes will allow LibraryThing to grow far beyond its current 1,498,000 books.

Although theoretically simple, the changes need to be made dozens of places. And tested. I’m behind on the sleep department, so I may not be able to finish it this evening. It will almost certainly be up by 5pm tomorrow, spiffier than ever. Within an hour or so I expect LibraryThing will surpass both 1.5 million books and 2 million tags.

Thanks to everyone who send letters of sympathy, encouragement and offers of assistance. Incidentally, I love the idea for the down page to sport a falllen pile repair-themed books.

PS: Is it just lack of sleep that makes the phrase downtime update seem linguistically funny?

I’m taking LibraryThing down for an hour or so this morning, when you should be asleep anyway—you’ve got work tomorrow. I’ve made some changes that improve the speed issues; I hope this proves another step in the right direction.