Peter Morville (see Semantic Studios) is giving the keynote on day 3 of KMWorld & Intranets 2008.

Peter’s main argument: search is a critical part of knowledge management. To do it right, we have to focus on two things at the same time … the details, and the big picture.

Peter’s background … library and information science, faculty at University of Michigan, and author of various books, eg, Ambient Findability (2006), Search Patterns (2009).

On Information Architecture
– Information architecture … more than about organization, also about controlled vocab, taxonomy, more than the Web, findability. Balance of art and science. Requires creativity and risk taking.
– provide multiple paths to get to the same information. Eg, multiple navigation starting points.
– information architecture requires a good understanding of user experience.
– “usability” is important, but the term is over-used and little-understood.
– one of Peter’s contribution is the value honeycomb (see below) that teases value and usability apart. Eg, usefulness, desirability, accessibility, and more. “Valuable” is in the middle to remind everyone of the need to drive business value.

– can look at each of the honeycomb areas in isolation (“let’s do a credibility audit”) and in relation to each other.
– Google and Wikipedia have a very intimate relationship in searches and knowledge participation. Google will point to Wikipedia content high in search results, and people contribute to Wikipedia because they know it will be found. Enterprises need to strive for a similar symbiosis.

Case Study: National Cancer Institute
The National Cancer Institute (see www.cancer.gov) wanted to improve its ability to provide information to cancer sufferers. Peter needed to argue that the role of the Web site was to do more that provide great information, but actually to go beyond and make it more widely findable.

Enterprise Findability
– AIIM market research study … really bad findability in the enterprise.
– a Fortune 500 company (“that shall remain nameless”) had very bad findability systems internally
– enterprise findability is IA + KM + Search. In collaboration, IA is emergent and there is a need to observe, shepherd and harness the learning to make things navigable, searchable, etc. Search needs to work across multiple systems

Looking into the Future
– information architects need to look to the past (to learn from lessons) and to the future (our systems will be in place for a long time).
– “findability” … the ability to find anyone or anything from anywhere at anytime. There are a lot of privacy issues with this.
– “chain libraries” … in the Middle Ages, books were chained to library desks to stop them from being lost.
– “a wealth of information creates a poverty of attention” (Herbert Simon)
– how is the wealth of information impacting on our ability to make decisions.
– Peter talked about Ambient Devices, and some of the capabilities to alert people to information things, eg, new email.
– “the Internet of objects” … the ability to find physical objects that have been tagged, and to display location on the screen. Eg, a hospital saving $28,000 per month trying to find hospital wheelchairs, by tagging wheelchairs.
– see The Transparent Society for a great book on privacy.
– in a world of lots of information, how do we create much bigger needles

– metadata has become central to the blogosphere … and it’s a great idea, but it’s not enough. Needs to meet tradition approaches, like IA and taxonomy.
– the “future of findability” … in 5-10 years, we’ll still be using a search box as the main place to start.
– search is iterative, interactive and a process of learning. See Marcia Bates (1989).

– search involves looking at a whole set of issues … who the user is, what are they looking for, defining content policies, search result interfaces (critical when users get stuck).
– see Peter’s Flickr photostream on search patterns (primary research for his 2009 book on this)
– some behavior patterns, eg, narrowing, search & browse & ask, and pearl grow.
– quite a few design patterns, eg:
– … best bets (a few good starting points suggested by humans)
– … federated search (users don’t know where to look, searching based on topic and not held apart by format)
– … faceted navigation (multiple ways to search and browse in combination, deals with the need for narrowing, provides a map of the search results)
– … auto-suggest (guess ahead on starting search strings)
– … structured results (eg, see how Google does this below)

– … social search (showing the inter-relationship between content and people; IBM is doing great work in this area on its W3 intranet, especially by giving a preview)
– … media search (for images and video, by color and shape and pattern)
– … spime search (eg, for searching real world objects in a shop)
– … redefining search (think Google Books, “expanding the searchable web”) (creating links between physical objects and the web)

Exploring Possible Futures
– there are a huge range of possible futures in search … with possible design ideas, and many different design tools.
– key final statement: search is a wicked problem … lots of uncertainties and incompletes. The problem in never solved.