Gilbane, San Francisco, 2008

I spoke at the Gilbane Conference in San Francisco in June (http://gilbanesf.com/) on the subject of using folksonomies as a publishers tool for classifying content. The session moderator had warned me that attendance was a little thin and that we shouldn't expect too many in the audience, but my rough estimate was that we had about 45 people there. I was pleased. And there were lots of questions and a lot of discussion.

Here's the abstract from my presentation:

Folksonomies, Just Good Enough For All Kinds of Things
Extending Folksonomies to Describe All Kinds of Content

Folksonomies are gaining popularity in the content delivery world as a "good enough" way to classify content; tag clouds are becoming commonplace on content-driven website. More and more, folksonomies are working their way back into the content creation and management process. Huge volumes of content, increasing and diversifying user requirements are pushing content management operations to look to folksonomies as a Web-2.0 way to describe the "aboutness" their content.

With this simple concept in hand, why not use this technology to describe more than just content subjects? "Aboutness" is just one aspect of your content. What about order, threads, or related content? This talk will explore the extended use of folksonomies as a technology for enhancing content in new and different ways.

I can send the power point to anyone interested, please email me or leave a comment on this blog.

Using Folksonomies in Content

I've been thinking more and more about folksonomies as a replacement for traditional taxonomies. We can create all the tools we need, or work off of existing tools such as del.icio.us. But in the end, size does matter. For a folksonomy to work, a *lot* of people have to look at and tag the content. To me, this means content has to be exposed to readers, and lots of them, to tag the content.

Traditionally, the classification operation has been something done under control and part of back-room content management, by a small, select group of indexers. So first, we have to decide to relinquish some control. While this may seem scary, in reality the volume of tagging makes up for the lack of specific control. And, we can build the tagging tools so that editors and indexers can follow behind the tagging and clean it up. But if we can achieve a large volume of tagging, the volume and repetitive nature of the tagging will create common tags, threads and relationships.

Second, we need to find a large group of taggers. Depending on the nature of the content, this can be accomplished in a couple of ways. First, and perhaps easiest, is to expose content to the web. Perhaps through incentives, taggers can be enticed to tag. And, of course, staff of the publisher should be encouraged to participate as well. Failing that, a publisher could look at a human-automation engine such as Amazon's Mechanical Turk, where large numbers of minuscule tasks that are best done by people are spread out over many people for a small fee.

Folksonomies Applied

In a previous blog entry, I was starting to think about folksonomies as they might apply to content management. Having lots of experience in topical classification, therauri, and related tools, I figured this would be a simple discussion, but it isn't.

What is a Folksonomy? "Folksonomy (also known as collaborative tagging, social classification, social indexing, and social tagging) is the practice and method of collaboratively creating and managing tags to annotate and categorize content. In contrast to traditional subject indexing, metadata is not only generated by experts but also by creators and consumers of the content. Usually, freely chosen keywords are used instead of a controlled vocabulary" (Wikipedia).

By their nature, folksonomies are created by the people. I think the key to a successful folksonomy is participation by many people -- we make up for the lack of a controlled, standardized vocabulary and it's application to content by sheer volume and enthusiasm from a wide-spread using community. In fact, complaints about this approach usually center around the imprecise nature of the tagging. Since users typically apply tags to content, the tags are often ambiguous, overly personalised and inexact. But, Guy and Tonkin make a persuasive argument that user applied tags are in fact converging -- that the overall universe of applied tags is becoming self-limiting. If so, then the universe of tags that are being created in services like Flikr and Del.icio.us are becoming useful bases for classifications.

And in other, related developments, these services are beginning to categorize their tags. Especially at Del.icio.us where there are now classification tags and action tags, among others. To me these seem like we gray-beards call facets.

But that's not really my point, though I think it is important. I think we need some tools that can work with content management to allow tagging, maybe even super tagging, wherein the tags are members of controlled facets. This isn't really hard. Virtually every content management system "knows" about content by a URI. And there are some very cool features in the Del.icio.us service, including keeping my bookmarks and tags private, and retrieving them via an API later. So, we can set a bookmark in Del.icio.us containing the URI of the content we want to tag with the tags we want for that content.

I tried this very simply by registering at Del.icio.us, turning on the "private bookmarks" setting, and putting the Del.icio.us buttons on my browser toolbar. Then I pointed my browser at a content item in a CMS (Alfresco) and clicked the Del.icio.us "Tag" button. Added tags and saved it. I can see the tags and the URIs in my items on the Del.icio.us website.

But, to actually use this information, we need to pull the tags and content URIs back out of Del.icio.us. XML to the rescue. Or, rather, XML and the Del.icio.us API. We can fetch the tags we're using by using this URL: https://api.del.icio.us/v1/tags/get. We can also see all of our content using this URL: https://api.del.icio.us/v1/posts/get. And we can retrieve by tag: https://api.del.icio.us/v1/posts/get?tag=C

Pretty cool stuff.

A New Kind of Content Management

I've been enamored with classification of content through my entire career (which is certainly long enough). Early on, I was working on library automation systems and in particular with Library of Congress Subject Headings (LCSH) data and I always thought classification of documents and other content types was a panacea we should try to achieve. Librarians used these and other tools as finding aids to locating specific information, usually books.

Along came the internet and the amount of information available grew exponentially. Search engines provide keyword access to text-base content. Publishing was redefined to include anyone putting information out on the internet. Published content became a thing to be managed with buzzwords like single-sourcing, XML, multimedia, syndication, blogs, Wikis, and so on.

Well, the world is very different now and with everything from books and documentation to audio and video as well as thought streams all published on the web, finding a particular bit of information can be a nightmare. Even traditional textual content is hard to find in the vast archive that is the internet.

So how do we find things now? What are our modern finding aids starting to look like?

Some very cool new technologies have come along recently: folksonomies and mashups, to name two.

Folksonomies in particular offer a new approach to an old problem -- how to capture the "aboutness" of content. Classifying content is an age-old process that has spawned many fields including taxonomies and indexing. These, in turn, have created an arcane set of rules and procedures wherein the maintenance of the indexes or thesauri become big efforts on their own -- at times larger than the content management effort they are supporting.

Mashups are a cool way to combine content from multiple sources into a single presentation. Kind of like federated search and portals in a social-networking environment.

I'll be exploring these more over the next several posts.