Speaking Thursday at the Boston Library Consortium’s annual meeting in the beautiful Boston Public Library, my focus was on the status of our library systems and the importance of remixability.

My blog post on remixability probably covers the material best, but my slides are online as both an animated QuickTime and PDF.

BPL, BLC, boston library consortium, boston public library, presentation, remixability, library, libraries, library systems, l2, lib20, library 2.0

I certainly don’t mean this to be as snarky as it’s about to come out, but I love the fact that Isaak questions my claim that linkability is essential to online discussions (and thus, communities) with a link:

• Linkability Fertilizes Online Communities
  I really don’t know how linkability will build communities. But we really need to work on building support platforms for the public to interact with the library and promote social discussions, whether offline or online. Currently, the only way for such interactions is through the High Browse Online blog, but even then, there are not much discussion going on. Maybe librarians need to go in more often to post useful comments. And we might need to read this to find out how to solicit more comments on the blog. [link to High Browse blog added -- Casey]

The big challenge here is to realize that the entire web is an online community, unbounded by geography or even the narrow confines of a single website. And I’m not sure I could point to a better example of that than Isaak’s post. Rather than comment at my blog, he instead posted on his — with a link back to mine.

But hey, you’ve gotta love that they’re hosting rock shows in the library.

book discussions, book talk, community, conversations, durable link, findability, lib20, libraries, library, library 2.0, library systems, linkability, online community, permalink, social software

I was honored to join the conversation yesterday for the latest Talis Library 2.0 Gang podcast, this one on folksonomies and tags. The MP3 is already posted and, as usual, it makes me wonder if I really sound like that. Still, listen to the other participants, they had some great things to say and made it a smart discussion.

I approached the conversation with the notion that what we were really talking about was whether libraries should give their patrons the opportunity to organize the resources they value in ways that make sense to them. For some time one of our patrons here has been telling us he wants all the books that he’s interested on one shelf, and for years the standard retort has been a chuckle. But, why, he might today ask, can’t our library systems make this possible in some virtual way now?

Tags — specifically user contributed tags — are a big element in this larger question. Though they bring up all manner of concerns from authority to vocabulary control, they’ve shown great value outside libraries and interest in them has been energized with the active discussions about how to re-imagine our library catalogs for today’s needs.

My big question is who “owns” those tags, and what motivates taggers. LibraryThing, has enjoyed some great success with tags, while Amazon has achieved little. Tim Spalding’s theory on the matter echos Josh Porter’s dissection of “The Del.icio.us Lesson,” where he notes that “personal value precedes network value.” That is, people tag for personal, perhaps even selfish reasons. People don’t tag to help the community, they tag because it helps the tagger.

I’ve been tagging my stories at MaisonBisson for some time now, and the effort has paid off by making my content more findable both internally and externally at services like Technorati. Flickr makes tagging even more valuable, as the tags are often the only searchable content for image. How else could I find my library-related photos if not from the tags?

On the other hand, my own experiment in user contributed tags seems to have fallen flat, as I’ve received very few useful tags despite the high number of readers who I’d expect to be familiar with tagging. Meanwhile, del.icio.us tells me that there are 133 tagged bookmarks to MaisonBisson in their database. This leaves me wondering if I should invest more effort in working on user contributed tag system that lives in my blog (or or a library catalog, or other discrete system), or should I instead focus on making those systems support outside tagging systems like del.icio.us? This is easy for my blog, where all the pages are already URL addressable, but bibliographic systems are a bigger challenge.

update: hey, Abby’s talking about this over at Thingology and her headline is way better than mine. Darn. Still, the point remains: we need to leverage our systems to make it easy for our patrons organize the things they like wherever and however they wish. Then, we should look for ways to find value in the aggregate. That’s the del.icio.us lesson.

folksonomies, folksonomy, interview, l2, lib20, libraries, library, library 2.0, library catalogs, library systems, opacs, podcast, tagging, tags, talis, talking with talis

“Bagged products” is little better than “cookery.” I’m gonna bet that no customer has ever asked the sales people for “bagged products,” that nobody’s ever checked the yellow pages for “bagged products,” and without context, nobody would come close to answering a question on what the heck “bagged products” are all about.

But we do have context.

And within that context, those two words are probably meaningful enough to the potential customers driving by. “Nursery stock,” “pavers,” and “bagged products” are just a few facets of that potential customer’s search for “landscaping” or “gardening supplies.”

The challenge here isn’t to reinvent our vocabularies, but to build systems that help the user who searches for “cookbooks” find more of them without needing an MLS to know the specific terms we used to catalog them. As it turns out, that search returns facets that give the user a hint that “cookery” might also be a good search term (it’s not perfect, but I’m happy to have any examples in this subject in my academic library to point to).

Aside: can somebody explain to me why a book might be cataloged as “Cookery, Indic” rather than “Cookery — Indic”? It’s not like “United States — History — 19th Century” would ever be represented as “United States, History — 19th Century” or “United States — History, 19th Century.” Or would it?

bagged products, language, categorization, subject assignment, classification, librarianship, libraries, lcsh, usability, findability, library systems, search, facet, facets, contextualized, contextualized results

It’s hard to know how Fuzzyfruit found the WPopac catalog page for A Baby Sister for Frances (though it is ranked fifth in a Google search for the title), but what matters is that she did find it, and she was able to link to it by simply copying the URL from her browser’s location bar.

The link appears among her comments in the discussion about her post on an early letter she’d written to her mom. Fuzzyfruit’s comment spawned more comments about the book from Sarahq and Coffeechica.

We talk here and there about how “libraries build community,” but how does that work in the online world? How do our systems support or inhibit community discussions online?

book discussions, book talk, community, durable link, findability, google economy, lib20, libraries, library, library 2.0, library systems, linkability, online community, permalink, social software

It was more than a year ago that Lorcan Dempsey pointed out this bit from The Chronicle:

Librarians should not assume that college students welcome their help in doing research online. The typical freshman assumes that she is already an expert user of the Internet, and her daily experience leads her to believe that she can get what she wants online without having to undergo a training program. Indeed, if she were to use her library’s Web site, with its dozens of user interfaces, search protocols, and limitations, she might with some justification conclude that it is the library, not her, that needs help understanding the nature of electronic information retrieval.

So I’m continually surprised to find people defending our old, broken search systems in the face of pressure from internet search services that we’ve all come to depend on.

Thing is, where’s the greater criticism in this?

Google’s simplicity and impressive search prowess trick students into thinking they are good all-around searchers, and when they fail in library searches, they are ashamed as well as confused.

Does the above serve to emphasize the “laziness” and “intellectual inferiority” that seem to be the theme of so many of our discussions about “today’s students?” Or does it instead show how backward we and our systems are?

I second Peter Binkley’s retort, but I also want to point out the huge wedge being driven between academic libraries and our patrons. We see it in the two perspectives on that quote above. Right now it’s a matter of not meeting patron needs, but I’m also worried about what happens when those students become faculty, and later, administrators. How will they value the library then?

(Title graphic from GoogleAlive. Go play, it’s fun.)

google, information, information behavior, information search and retrieval, lib20, libraries, library, library 2.0, library systems, online behavior, peter binkley, search behavior, search engines, search practice, Steven Cohen, web searching

I’d be excited just to be a fly on the wall at code4lib, but I’m on a bit of a mission to change the architecture of our library software — to make it more hackable, and make those hacks more sharable — so I had to propose a talk.

Title: What Blog Applications Can Teach Us About Library Software Architecture

Description: The number of programmers in the library world is growing and our individual efforts have shown great promise, but they exist largely as a spectacle that few libraries can enjoy. We need better means to aggregate our efforts and share solutions that can be employed by libraries without programming staff.

We must look outside libraries. The blog world is rich with non-programmers using (maintaining and configuring) blog applications like WordPress or Moveable Type (or others) while thousands of developers are adding functionality via plugins and themes. What lessons can we learn from this and how might an OPAC built from those lessons work?

The idea here is that lots of people are blogging, and a good number of them have figured out how to install and maintain their blog management software — even installing plugins and tweaking templates. What lessons can we, as a community of library programmers, learn from that?

Yes, I’m working on a project along these lines, but it’s not yet ready for public consumption. You’ll see more about this in future posts, and if you’ve been watching, you might have picked up on it in previous posts.

library, libraries, library 2.0, library20, lib20, code4lib, system architecture, reusable code, common platform, system design, software design, software architecture, library software, library systems

I’d meant to point out these two articles from Library Journal ages ago, but now that I’m putting together my presentations for next week (NEASIS&T & NELINET), I realized I hadn’t.

Roy Tennant writes in Doing Data Differently that “our rich collections of metadata are underused.” While Roland Dietz & Carl Grant, in the same issue, bemoan the dis-integrated world of library systems.

tags: dis-integration, integration, libraries, library journal, library systems, metadata, using metadata

Peter Morville, author of Ambient Findability, stirred up the web4lib email list with a message about Authority and Findability. His message is about how services like Wikipedia and Google are changing our global information architecture and the meaning of “authority.”

The reaction was quick, and largely critical, but good argument tests our thinking and weeds the gardens of our mind. Argument is good. Here’s my side.

It’s important that we understand how modern search engines work. On the web, each link is a citation, and citation analysis is an important component among the many algorithms used to rank search results. Highly ranked content appears at the top because it is frequently cited (linked). This is obvious to many, but what is harder to fathom is that we (those who publish web content, anyway), not the search engines are responsible for identifying value on the web. Each link is a value statement about the resource we link to.

Think about that in the context of this comment:

Just because a document is findable, this does not mean that its contents are *better* or more truthful than a document that is not findable.

My point is that findability is in fact a measure of value. A perhaps incomplete and indirect measure, but one that has shown a remarkable ability to deliver valuable and useful information on demand.

Unfortunately, not all content is available online, and not all online content is linkable. Sadly, many web OPAC pages are not linkable, as is true of most every A&I and full-text database (or the content is linkable but inaccessible behind an authwall).

So now this quote:

As librarians, we are supposed to be experts on helping people find and retrieve quality information.

And I have to follow this with the following question: How better to help our patrons find high quality, accurate, and authoritative information than to take advantage of the search engines that already answer hundreds of millions of questions each day?

As stewards of knowledge, we need to understand the Google Economy. We need to build applications that embrace it. We need to invest the value that librarians bring to the search for knowledge in our online services.

Feh, libraries are full of people smarter than me. Hopefully they’ll forgive me for speaking out of turn.

tags: authority, citation analysis, findability, google, google economy, libraries, library, library systems, quality data, research methods, search, search engine, search engines, web opac, wikipedia

Ryan beat me to reporting on the interesting new services at the Ockham Network (noted in this Web4lib post). The easiest one to grok is this spelling service, but there are others that are cooler.

He also alerted me to a Perl script to proxy Z39.50 to RSS. Though for those more into PHP (like me), I’d like to point out the YAZ extension from the folks at Index Data.

This extension offers a PHP interface to the YAZ toolkit that implements the Z39.50 Protocol for Information Retrieval. With this extension you can easily implement a Z39.50 origin (client) that searches or scans Z39.50 targets (servers) in parallel.

The module hides most of the complexity of Z39.50 so it should be fairly easy to use. It supports persistent stateless connections very similar to those offered by the various RDB APIs that are available for PHP. This means that sessions are stateless but shared among users, thus saving the connect and initialize phase steps in most cases.

Clearly, if I wasn’t so lazy, I’d get around to compiling YAZ into my PHP and actually build something with it. Though without Z39.50 on my own ILS, the incentive is understandably diminished.

Finally, I’ve gotten an answer about indexing cat dates in III so they can be used in searches or sorts. The answer is mostly “no,” but I go into a little more detail at LibDev.

tags: cat date, ils, index data, information retrieval, libaries, library, library catalog, library systems, ockham, opac, perl script, php, rss, search by cat date, sort by cat date, xml, yaz, z39.50, z39.50 proxy

I’ve been looking seriously at metasearch/federated search products for libraries recently. After a lot of reading and a few demos I’ve got some complaints.

I’m surprised how vendors, even now, devote so much time demonstrating patron features that are neither used nor appreciated by any patrons without an MLS. Recent lessons (one, two, three) should have made it clear that libraries need to conform to patron expectations of how online resources should work. Our own search statistics show that only 0.0067% (YES, less than a hundredth of a percent!) of the searches on our OPAC get “limited” to specific languages, locations, dates, or material types. What our patrons expect is that a natural language search will yield relevant results in the first page of hits. “Googlization” isn’t about dumbing things down, it’s about making the technology smarter.

And that’s the problem with these vendor’s metasearch products. They don’t do much to improve the quality of the results retrieved from any database. Shovelware, products that pile up junk in an attempt to generate value based on quantity, is poor solution for libraries or researchers. Still, that’s how these products work, and it’s how they’ll continue to work until libraries and their database providers adopt some of the advances in search technology now used on the web (it’s not just Google, but Yahoo, Teoma, Clusty, and others).

At the same time these metasearch products are doing little to improve the results we get, they’re also making the search process slower. Why do they all make us wait while slowly updating a table that shows only the number of hits retrieved from each database? A9 can teach all these vendors quite a few lessons on that point. A9 reports results in resizable columns, and fills in the details from various databases as they become available. The biggest lesson A9 can teach these vendors, however, is that metasearch should be free. They’re pushing OpenSearch as a public standard based on RSS/XML and already they’ve got access to 236 databases. That’s not bad compared to z39.50 (which we all still respect as the elder parent of current search standards), but remember that the standard was only announced in March 2005.

We need to pressure database vendors to improve their search engines and give better results. Maybe database providers need to rank journal articles by the number of citations they receive? Maybe libraries need to buy Google Search Appliances and do their own indexing of database content. That way, links from university faculty would increase the rank of articles they link to, making search results especially relevant.

Then, we need to ask where our money is going when we buy software like this. We need to demand standards-based products with outstanding ease of use. Go try out A9 and compare it to anything in your library. Yeah, don’t you wish you could offer that to your patrons?

tags: a9, clusty, federated search, google, libraries, library, library systems, metasearch, natural language search, opac, patron, patrons, search technology, teoma, yahoo