I've been pondering this for a few days and wanted to put it out there for feedback. It's full of holes and fairy dust, but I'd love to hear what people think.

I was looking at an RDF graph visualization the other day - colorful linked clusters, like rigid versions of the pom poms I made as a kid. Clusters of related data, with occasional links to other clusters of related data, and other clusters nearby. Visual Complexity has a wonderful set of images representing data in this fashion. Reminded of why we librarians labor to determine the best shelving locations for books - so that similar topics can be scanned easily.

[For now we'll put aside the small public library that - rightly - shelves it's majority fiction collection alphabetically by author.]

Determining just the right order to shelve those books in? A royal pain. Keeping objects on the shelf in the right order - never happens if you let real people touch them. The magical way we do this - it's usually a technique called Cuttering. Cuttering is supposed to keep the objects on a shelf in alphabetical order with a unique identifier. Cutter's are meant to be thought of as decimal numbers, even though there's a leading alphabetical character and usually no decimal point in the number. An example of a Cutter number is here in bold: HD30.2 .L36 2007. Every library gets to make up it's own local rules on how to create Cutter numbers, yay, but are supposed to base them on alpha-to-numeric mapping tables from one of several revisions over the last century. The mappings are not nice and easy though - there are only 10 nine digits and 26 letters in the character set Charles Cutter based his work in. Most libraries base the Cutter number creation on the last name of the author, but there are extra tables for genres, geographic materials, music, translations, and who knows what else. I can't do this justice - go see Greg Cotton's course page. With all due respect to Professor Cotton - and please note, I received my library degree elsewhere - this is one of the reasons I HATED my required Cataloging and Classification course in grad school. MARC, Cuttering, AACR2 - ugh.

So what if we were to abandon the notion of straight lines of shelving, all the books nicely ordered alphanumerically? What if we could have round or square or pentagonal shelves, on which related items were kept, where facing units could have some kind of relationship? In theory it could work. Much like how I order my own collection of books. In most homes, neighboring rooms have a relationship - the kitchen is next to the dining room, bedrooms or personal spaces are clustered. I scattered my books around the house - cookbooks in the kitchen, art books on coffee tables, business and computing books in the office area, craft books in the craft area, fiction and reference on the library shelves. The kids books go in each of their rooms - my own version of social tagging of my books; after all, I had to give each child a copy of Good Night, Moon and I know which copy belongs to which child by sight.

We maintain the ability to browse a topic. We can easily turn and see related topics. We leave room for welcoming comfy chairs to enjoy our finds. With appropriate signage - real world Information Architecture - finding the right cluster should be no more difficult than find the shelf containing the P325 sub-class of the Library of Congress Classification scheme. We lose the task of Cuttering and expand the ROI of applying subject headings.

OK, I haven't exactly figured out how to determine exact shelf order. The OPAC doesn't care if the shelves are straight, it just needs to tell the user where to go. The librarian may have to get used to non-linear spaces, but they are more than capable of telling a user where to find useful or desired information in any situation. And you may wonder about instance identification, but frankly, many libraries today use the practical convention of "Copy 1," "Copy 2" or "c1" or "c.1" to indicate instance. What's to say we can't use RFID tags or another unique identifier? And what of objects that belong in more than one cluster?

This is what happens when I wake up in the middle of the night, my brain randomly piecing bits together. I hope I've either amused or inspired someone out there! I'll keep poking at it. :)