Donut Age: America's Donut Magazine

One LibraryThing to rule them all...

Sometime early this morning, LibraryThing surpassed 1,000,000 catalog entries (they were at 996,000 before I went to bed). That's pretty remarkable growth for a site that only opened three months ago. I've been doing my small part, having cataloged over 500 of my own books in the couple weeks that I've known about the site. I'm small potatoes, though; there are users with libraries in the thousands (I can only assume they had some pre-existing catalog of their stuff and are taking advantage of the bulk-import feature).

As previously mentioned, LT has major addiction potential for book nuts, into which category I obviously fall. The social aspect of seeing who else owns the same books as I do and what they think of them is fairly entertaining. I've also been spending inordinate amounts of time looking at my Tag and Author Clouds as visualizations of my own interests. And I think the rating and reviewing features are going to jumpstart me into semi-systematically posting some book reviews here (so I can link to them from LT).

Using LibraryThing this intensely, I find myself frequently weighing it against my favorite piece of social software, That may not be entirely fair since has been alive and maturing for over two years while LT is still in its infancy and has been changing rapidly. Nevertheless, the comparison is too tempting to pass up.

What's perhaps most noticeable is that I've spent a fair amount of time worrying about the information I put into LT — about getting it "right" — whereas one of the main attractions of for me is that I don't worry about that. I think this has to do with what I catalog on each site and why. On, I am cataloging links for the purpose (primarily) of retrieving them later. For that, there's really only one essential piece of information: the address. There's lots of other things I might want to know about a website, but the URL is both a unique (mostly) identifier of and sufficient means of retrieving the site. This allows me to be pretty sloppy with tags and other information; I am just trying to make sure I can find the bookmark again when I want it, so whatever will help me remember it s good enough.

The general purpose of LibraryThing is different. Retrieval is not really a concern; LT assumes you already own the books you are entering and know where to find them. I have seen people add tags like "at office" or "bedroom" to their entries, and I suppose if you have 8,000 books in your library you might need such notes, but even then, there's a difference. A bookmark on links directly to the website in question (as long as it still exists); location information in LT is just information, it doesn't provide access to the book itself. You can use LT to maintain a wishlist (it includes links to Amazon and other booksellers), which is closer to the way I use, but that's not the main impetus behind the service.

LibraryThing focuses instead on gathering descriptive information about your books and using this information to connect you to other users (by identifying similar libraries, making book recommendations, and through book and author pages). However, cataloging books, as generations of librarians will attest, can be a complicated affair. One central dilemma has to do with what constitutes "the same" book. Many of LT's social features hinge on identifying users who have "the same" books in their libraries, but "the same" means different things in different contexts. For many purposes, the same author and the same title add up to the same book (what textual scholars would call a "work"). You read Homer's Odyssey, and I read Homer's Odyssey — we have something in common so connect us! But wait, not everyone will agree that all Odysseys are the same work. Classicists would certainly not want their scholarly Greek text being equated with my English translation. I despise verse translations for the most part, so I don't want your Fitzgerald getting mixed up with my Lattimore. For even more specialized purposes, the particular "text" or instance of a work becomes important. Joyceans have heated debates about different editions of Ulysses (I have four on my shelf: the 1961 Random House edition, Hans Gabler's 3-volume critical & synoptic edition [1984] and the corresponding single-volume trade paperback [1986], and a 1993 Oxford edition that may or may not reproduce the original 1922 text). Pound scholars will sometimes make note of the printing of the Cantos that they use as there are subtle differences among these. Collectors and bibliophiles will care about hardback vs. paperback and other paratextual features. In some cases, there are even unique copies that deserve distinction (as in a signed or annotated volume).

Librarians and bibliographers have developed systematic ways of describing books on all these different levels. The difficulty comes in deciding how far to go because what is crucial information for some users is irrelevant or even confusing for others (LibraryThing's creator, Tim Spalding, has discussed the difficulty of making this decision on his blog). LT compromises by incorporating some traditional library catalog fields like author, title, date, and publication information while also allowing tagging and a free-form Comments field. It's not a bad compromise, and it allows some some added capabilities (I like the "author cloud" feature), but the formal taxonomies tend to de-emphasize the flexible "folksonomy" of tags. It is telling that the full list of one's tags is kept on a separate page from one's actual catalog in LT. In, tags are omnipresent and do almost all of the work. The tag list or cloud on the side of almost any page functions as a navigation tool. The ability to combine tags serves as a search mechanism (e.g., system:filetype:mp3+music+jazz). The special for: tag allows you to recommend sites to specified users. Having gotten used to this kind of power, I find the comparatively limited tag uses in LT somewhat restrictive. I also find myself thinking of LT tags in stricter terms that tags, fretting, for example, about whether 'medieval' should be used only for books from the medieval period or also for books about that period. Perhaps the latter is just me clinging to hierarchies I've habitually associated with books, but it is striking how much more anxious I am about my LT entries than my ones.

None of this should be construed as saying that I'm unhappy with LibraryThing. On the contrary, I think it's marvelous. But LT faces some peculiar challenges in the way it organizes information, and it will interesting to see how it addresses them as it matures. While folksonomy and social software are all the rage right now and certainly come together beautifully on, they may not be the perfect solution for all things.