Raising Arizona

By Karen G. Schneider |  

What do two initiatives from a new branch library and a large city library have in common? Both--from Maricopa (Az.) County Library District and Phoenix Public Library--are about moving out of library silos and leveraging the wide world of information.

Years of focus groups had taught the Maricopa County system that 80 percent of their users came to their libraries to browse popular reading, and Dewey organization didn't meet their needs: it wasn't friendly, and it wasn't familiar. Complaints from users indicated they wanted the library to be more like a bookstore. The new Perry branch was an ideal test bed for rethinking organization, because the community hadn't been conditioned into what to expect in a neighborhood library and also because Maricopa would stock the library with a brand-new, opening-day collection, so that there weren't issues with retrofitting records.

So Maricopa elected to organize books by BISAC headings. BISAC — for Book Industry Standards and Communications— comes from the bookselling world, and is "an industry-approved list of subject descriptors." BISAC has 50 major descriptors and many more subheadings. You've seen BISAC headings on shelf labels if you've ever walked into Borders or Barnes & Noble and seen shelf labels such as "Computers," "History," or "True Crime." 

Library staff worked with vendors from Brodart and Baker & Taylor to select the right BISAC subheadings, and the library, had a soft launch in June and officially opened on July 7. (See these photos from "Gather no Dust" blogger Jeff Scott's site visit.)

The result? Circulation is robust, patrons are happy, and though the library staff steeled themselves for "outrage," not a single customer has complained. As Maricopa County library director Harry Courtwright points out, if it doesn't work, they can always go back to Dewey--but they are optimistic enough, based on the initial response, that they are planning to use BISAC organization for a library they are opening in 2008.

As the mainstream press observed, there was a fair amount of wurra-wurra on some library discussion lists, with librarians asking why a library would abandon something as "proven" as Dewey, and others predicting the end of the world as we know it. Librarians also struggled to argue that sure, using BISAC would work for a small neighborhood library branch, but it could never, ever work in a larger setting.

Some librarians--in what sounds an awful lot like sour grapes--said, sure, fine, but this wasn't really innovative; it had been done elsewhere. To which I would observe that in the marketplace of ideas, the race often goes to the strongest, clearest voice. Bold theft of ideas from other venues is just another innovative strategy.

Meanwhile, other librarians wondered aloud about the "additional expense," unaware that Maricopa buys all of its books precataloged, and that vendors assign BISAC to books as a matter of course.

The Plot Thickens

Meanwhile, over at Phoenix Public Library, Jesse Haro in the automation department has been demonstrating that at least in the world of online library catalogs, BISAC may be better than Dewey for topical browsing of large library collections. (Oddly, Phoenix Public Library is in the same county as Maricopa, but in the Byzantine world of library organization, Maricopa is its own tax district and Phoenix is a city library.)

Haro and crew initially set out to improve their OPAC by replacing it with Endeca, which offered distinct improvements such as more relevant search results, spell-check, and the ability to integrate tagging, pick lists, RSS feeds, and much more.

However, Haro encountered a problem I discussed in an earlier article about NCSU's implementation of Endeca. I commented that while adding facets (guided navigation) to the OPAC was a huge plus, in the end, the usefulness of the facets was limited by the browsing language used to generate them, and I added that Library of Congress Subject Headings are "not designed for browsing collections on the Web."

I didn't mention Dewey, but that system was in my mind as well.

BISAC codes are pragmatically user-centric; they're designed to make it easy for customers to browse for books. The language is simple, the subcategories broad, and the main groupings are designed around user browsing and buying habits, such as "I'm looking for new mysteries" or "I am planning a wedding."

Melvil Dewey, on the other hand — and this is no criticism of his efforts — was designing middleware. Dewey invented his Decimal System in an era when most libraries had closed stacks and library workers retrieved known items or conducted their own "browsing" on behalf of patrons. The Dewey system wasn't designed to be easy for casual users in a neighborhood library where the emphasis is on self-service; it was designed to be efficient for large collections organized and managed by knowledge workers.

Pouring a nineteenth-century inventory system into a twenty-first-century search engine can lead to--shall we say--interesting results. Haro comments that "Endeca exposed our catalog in ways that were for better and for worse." Once of the "worse" ways was the clunkyness of library-generated metadata for topic browsing.

Purely unscientifically--but based on five years working work with thesauri for web portals--my conjecture is that for browsing purposes (in either the physical or the online world), most classification systems are at once too broad at the top level, such as Dewey's nine categories, and too narrow at the next level, where our systems suddenly drop into deep water (and our classification systems are hopeless with fiction, for which many libraries have created their own codes). I am particularly familiar with the
"onesie syndrome," where a broad subject heading, when subdivided, suddenly leads to exactly one item (q.v. "Gorillas -- Congresses"). This is not a browse-worthy outcome.

In any event, as a profession, we've had notoriously little luck with teaching people to think "341.5" (uh, make that 641.5!) instead of "cookbook," or "917" for "travel."  So how did Haro and others at Phoenix decide to use BISAC topics? They went to Barnes and Noble and looked at how the bookstore organized its books -- exactly what Maricopa did -- and that's when the idea of BISAC came up.

Unlike Maricopa, with its brand-new collection, Phoenix had the problem of retrofitting existing catalog records. So catalogers at Phoenix enriched MARC records with BISAC headings (in the 695 field, as a record display shows), using a Dewey/BISAC crosswalk provided by the vendors. Endeca (along with other faceted-navigation search engines such as FAST and Siderean) can display any metadata as a facet—so now users can browse for books in the Phoenix Public Library catalog the way they can browse for books in the Perry Branch.

Courtwright says he has a team at Maricopa looking at their catalog, to see if the bookstore-style organization that's working so well at the Perry branch can be extended to their virtual display. I keep hoping the two libraries could at least have a coffee klatch to discuss their amazingly similar and delightfully innovative work.