Thanks for all the comments on my last post. It’s thrilling to see so much passion and intelligence applied to these issues. I’d like to experiement with continuing this and other deeper conversation at the recently-created LibraryThing group on Google Groups ( http://groups.google.com/group/Librarything ).

I’m still making up my mind about this group. In the long term I’d like the conversations to happen on a LibraryThing forum, and until then I still enjoy the blog-comments model, despite its drawbacks. But I’m game to see how this might work in a forum. You can, of course, continue to post on the blog too.

Here’s my kickoff message there:

Many of them seem to have feared that LibraryThing was in fact going to veer into Friendster-style social software. Nothing could be further from the truth! I am of the opinion that LibraryThing must always be a great way to catalog books, whether a user wants to catalog everything or just keep track of their reading list. The social aspects must be rooted in this.

Just because they’re rooted does not mean that the social aspects are unimportant or even—I would argue—secondary. Lots of LibraryThing’s features are “social” without being Friendster-ish. Over 2,000 bloggers have mentioned the site, most inviting friends to come look at their collection. By contrast, the Mac based cataloging application and cheek-pinching darling of the technology press, Delicious Monster, has less than 40% as many mentions during a longer period. Why? Because it’s no fun talking about something nobody else can see! Among these bloggers over 650 have added a LibraryThing widget to their sidebar. And all you guys—even if you argued against social features, took the time to post a message on the LibraryThing blog. These are social uses.

Or take the author or tag pages, a very “rooted” social feature. Without LibraryThing’s social machinery, these would be pretty boring. For recommendations you’d get what Delicious Monster and every other cataloging ap provides—the holy Amazon five. The data people put into LibraryThing opens up all sorts of interesting prospects. Take author disambiguation. You can’t do that on any other service. (And yes, edition disambiguation is on its way.)

I think LibraryThing’s strength lies in the unique combination of social and non-social features. Different users will appreciate different aspects. I know there are users who want nothing more than a cataloging ap, keep their library private and probably never click to find out who else owns their books. On the other end there are impatient people who catalog their ten favorite books, throw them up on their blog and look around for people to discuss them with. To succeed, I think LibraryThing needs to continue to appeal to both types.

I think the solution involves these principles:

• The book catalog should always be the central metaphor of the site.
• The site is both useful and fun without any social interaction whatsoever.
• Users control their social exposure (eg., private libraries, turning comments off).
• Most social features should emerge from book catalog data, not from photos, locations or non-book interests.
• Social features should be developed with taste and restraint.*

Thanks once again for thinking through these issues with me.

*The NYT recently had an article about the “eww” factor when Friendster premiered a feature that let you see who had been looking at your profile and when.

I’ve done some tweaks to the design. I’ve wanted to do this for some time and never found the time. The nail in the coffin was an email that described the top bar on the old design as “dangerously close to greige.” Ouch.

There’s a lot more I need to tweak, but that’s true of every aspect of the site. I’m worried that tweaks like this make LibraryThing look too corporate and static. But the old design was ugly and corporate and static, so I think it’s an improvement.

Playing with the design got me thinking about large issues of graphic and social design. How “accessible” should LibraryThing be? By accessible I mean what people mean when they call a book accessible—easy for a large group of people to get into. I’m a little afraid of making LibraryThing too accessible, too appealing.

User photos. One way of doing this would be to put user photos front and center. Social sites like Friendster, Tribe or MySpace, even LibraryThing’s (nearly moribund) competitor Mediachest do this. Photos add impact and—let us be frank—sex appeal.

Social software home pages are particular locus of user-photo activity, with the meat-market aspect paramount. I have no proof, but I believe most social software site’s “random pictures” algorithms have a one-hottie minimum. I particularly appreciate Mediachest’s non-rotating homepage hottie, “Becca from Seattle” (at right), who attests that she used Mediachest and “met some cute guys that actually shared my interests.”* As Twain says, I am girding up my loins to doubt this. Not only does the photo not reside in the user photo directory, but Mediachest requires a state and zip code, and allows a geographic lookup—no Becca in Seattle.

I think there’s reason to believe that LibraryThing users don’t want that sort of atmosphere. LibraryThing makes it very easy to add an image. (The preset ones are one of my favorite features. It warms my heart every time someone discovers the software “easter egg” there.) But of even serious users only about half have chosen one. Among the ones who have are the two lasses at right. Their favorite tags are “democratic capitalism” and “almost but not quite a dictionary.” Their favorite books are… okay, that’s from Friendster and appears to be a magazine photo. (Real people don’t wear matching underwear.) As far as I’m concerned, this is the enemy. I think most LibraryThing users will agree.

Friends. There are a lot of other places where LibraryThing could broaden its appeal and play up the social aspect. Another example is “friends.” At present LibraryThing allows you to add users to a “watch list,” not a “buddy list.” The difference is that nobody knows who’s on your watch list—it’s a glorified bookmark. Buddies lists are very public. I did this because I hate the social dynamics of buddies lists. “I’m his buddy but he’s not my buddy.” “She’s got 200 buddies,” etc. Maybe it’s just because I’m old enough to remember Reagan’s first term, but I find this sort of this pretty irritating and pointless. Am I being a stick in the mud?

Incidentally, I will be adding “groups.” I’m not quite sure how they’ll work but the idea will be to allow people to search for books within a set of libraries. The point is more functional than social. Users have been begging for a way for their whole knitting or book club to get on LibraryThing and search a combined library.

Wrap up. So where should LibraryThing’s design go? Would more use of user photos make the site more fun and immediate or do they lead down a slippely slope? Should I turn the “watch list” into a “buddy list.” Should LibraryThing allow users to enter zip codes and search for them, a la Mediachest? In sum, should I bend the design more to the conventions and expectations of other social software, or should LibraryThing try hard NOT to do this?

Comments appreciated. Sorry for the long post.

PS: Holy smokes—21,000 books added yesterday! That’s double the ususal rate. I think LibraryThing will hit 1.5 million books tomorrow, and 2 million tags the day after.

*Note: I am doing criticism, which entitles me to use the photo as fair use.

Update: I changed something so you can now use Universal Import to import BookCrossing books more easily. See the BookCrossing forum.

Quite a few LibraryThing users are also users of the innovative and long-running site BookCrossing. As the site puts it:

“BookCrossing lets you share your books with the world, and track their individual journeys forever more. Our members have registered 2,657,603 books so far, and a good number of these are “in the wild” at any one time.”

To make the system work, people give their books a “BookCrossing ID” (BCID), composed of three digits, a dash and nine digits. The first digits are secret—available only to people with their hands on the book. You need this part to log where the book has gone and make journal entries for it. The second set are public. Some BookCrossers enter all their books, even if they haven’t given them away.
I’ve added the BCID to the edit screen (the pencil, ). And you can add the field to your catalog screen (use the “change fields” feature). In your catalog you can click on the BCID to go to the journaling page. Needless to say, it only shows the secret part when you’re looking at your own catalog.

So that’s what it does now. I’m all ears. I don’t use BookCrossing, so I don’t know what other features would be helpful. Oh, and does anyone want to post about it there. I feel a little chintzy doing it myself.

Note: If you see the BCID you’ll also bought, started and read dates. I’m still adding support for this, but feel free to leave comments about how you want it to work or what data should be presented.

I’m looking for book-pile photographs representing Martin Luther King Day and Black History Month. Do you have a lot of books on the topic?

I did a general book pile photo contest earlier (see winners and all submissions on Flickr). It was fun. I want themed bookpiles to become a regular thing on LibraryThing, much the way Google’s logo changes for special occasions.

The rules:

• Take a photo of a pile of books. I favor piles with easy-to-remove backgrounds and legible titles. See previous winners for what I favored.
• Post it to Flickr tagged “librarything” and “blackhistorymonth.” If you must send me files, send them to timspalding gmail.com.
• Winner gets a free one-year gift subscription, to use or give to someone else.
• I’d love to close it in time for MLK day, but that may be too early.
• By participating LibraryThing gets rights to use the photo for promotional purposes, with acknowledgment.

Incidentally, if anyone wants to become LibraryThing’s Dennis Hwang, coming up with holiday book piles, like Hwang’s Google holiday logos, on a semi-regular basis, I’d love to hear from you. I can’t really afford to pay you, but I can showcase your work and give you free memerships from time to time.

UPDATE: Some speed issues remain but the crashing issue is mostly solved. I’ve fixed a number of things and added a few features. I’ll blog about them soon.

I just discovered that Neil Gaiman, author of American Gods, Anansi Boys and other works (see his LibraryThing author page), dropped a mention of LibraryThing on his blog. So that’s why twenty-nine visitors signed up last hour! Gaiman writes:

“Which I find, to be honest, an extremely attractive sort of website, and if I only had a month or so with nothing to do, I’d input my own books.”

I’ll bet people would love to know what’s on Gaiman’s shelves, and those of a lot of authors. LibraryThing has a few authors already, such as Ann Douglas (profile/author page), a baby and parenting author. It needs more. Best selling authors are hard to incentivize. Would Mr. Gaiman take a free account?

Aside: I shouted out the news to my wife in the other room. Neil Diamond’s blog? The singer? Not yet. Not yet.

I know LibraryThing has been slow. I am working to get it working faster.

I have the following plea. If LibraryThing does not respond quickly, please do not click “refresh” trying to get it to speed up. This never works and has been behind the recent downtimes. Things are going smoothly, the database server gets a little behind, and people start hitting refresh until the server is so behind it crashes.

Thank you for your patience as I work to scale LibraryThing to satisfy its current popularity.

1. Performance update. Performance has been spotty recently. Accelerating growth—Sunday saw 16,000 books cataloged!—has strained current resources. Improving speed and reliability is now my number one goal. Occasional slowness and even downtime will happen, but things are heading in the right direction.

I’m getting some part-time assistance. (But I’m still looking for a good LAMP hacker; email or send resumes!) So far I’ve re-jiggered settings and optimized requests. I am looking very carefully at the path up. I will probably be setting up a dedicated database machine and/or a dedicated “thinking” machine. I have a new book-suggestion algorithm that is to book suggestions what Deep Blue was to chess—extremely good and extremely processor-intensive. Bringing it live would crash the present site. It might also take over NORAD computers and threaten nuclear war until forced to learn the meaning of checkers.

I’ve temporarily disabled all-catalog searching while I work on database speed. These searches were taking up a large percentage of database power.

2. Two new features. Before I turned my attention to speed I did complete two new author suggestion features. Author pages now include a “Similarly tagged” box; tag pages sport include a “Related authors” box.

They are hit-or-miss. A decent example would Ilaria Gozzini Giacosa, author of the cookbook A Taste of Ancient Rome. With only a few tags the system managed to match her up the author of a Medieval cookbook and with Apicius, author of De re cocquinaria, the only extant example of a Roman cookbook. Apicius is in the unique position of being “similarly tagged” with both Marcus Tullius Cicero and Betty Crocker—both fusty and overrated? At the other end, take Nuala O’Faolain, author of Are You Somebody: The Accidental Memoir of a Dublin Woman. She is paired with Barack Obama. Both wrote memoirs, I guess. As more tags enter the system, this sort of single-tag effect should lessen. Here is the “similarly tagged” for David McCullough and the related authors for World War II.

Earlier on this blog I called for a Flash programmer. A blog reader found me one. (The new Flash-based widget will be coming out soon.) Now I’m looking for a LAMP guy—Linux, Apache, MySQL, PHP, particularly the MySQL part of that—to help me with some scaling issues. Is that you? Do you know her?

I have some limited and particular problems now (eg., the slow query log). But I could see this evolving into something more steady. I can pay a low-decent hourly, and you’d have the pleasure of working on something cool.

Bonus points for living in Portland, ME or Boston, MA. But I suppose this doesn’t matter these days.

Update: The widgets are back up again. They were responsible for a lot of the speed problems. I had no idea how often they were getting requested! Fortunately, the code had a lot of room for speed optimization.

I’ve improved the tag searching. You can now ask for things like:

• birds, beasts — all books tagged both birds and beasts
• +birds +beasts — another way to say the same thing
• +birds -beasts — all books tagged birds but not tagged beasts
• greek* — all book with tags starting with “greek” (eg., “greek”, “greek history,” etc.)
• * history — all books ending with “history” (eg., “greek history,” “roman history”)
• *dog*, -fiction — all books with the word “dog” somewhere in the tags, excluding fiction
  

Announcing a decent tag search for LibraryThing is a bit like announcing that I’m finally using deoderant under one arm. “What about the other arm?” Well, the regular search function is still quite suboptimal, but I’m working on it. I’ve got to fiddle with some internals to do it.

I’d love to hear people on the topic. It’s clear to me that there should be a link in the catalog view to search that catalog. That functionality is already there. But should the search tab do the same thing, displaying every matching book and user in the catalog view? This seems an admirable way of looking at your books, but pretty poor when looking at others. I’m thinking the search should go instead to a page like the tag or author pages, that “rolls up” all the copies under a single book title. Clicking on the book title on that page would take you to the book-detail pages, but there would also be a link to see the books in the catalog. Hmm…

LibraryThing has added 151,440 XML/RSS feeds, and an equal number of HTML versions of the feeds. Someone asked for “feeds for everything.” This is pretty close.

LibraryThing now offers five types of feeds:

• Recent books added by a member. Follow what a user is reading.
• Recent reviews by a member. Follow what a user is saying about books.
• Recent reviews of books a member owns. Find out what other people are saying about your books.
• Recent books tagged something. Follow new items for any tag in LibraryThing.
• Recent reviews of books tagged something. Follow reviews for any tag in LibraryThing.

As you probably know, RSS feeds look like . To work with them you need some sort of feed reader, such as Bloglines. You can also plug them into your Google homepage and many other “start pages.”

If you don’t use a feed reader—I don’t!—I’ve provided an HTML version for every RSS feed. I’ve given these links the icon .

At present feeds are updated every twenty minutes. If usage is high I may need to cut that back a bit.

How you can help. I know bloggers love RSS feeds, so I’m asking for your help spreading the word. (Some people never signed up because of this issue!) And, as usual, I’d love help debugging and expanding the feature. Is there anything I’m missing?

Also, the HTML version is something of a hack; I’m having a deuce of a time getting PHP5’s XSLT support to work. Charitable PHP hackers are invited to email and find out just how clueless I am about this sort of thing.