I recently attended the 2010 Enterprise Search Summit.  Here are my notes from the session, Search and Discovery Patterns, was lead by Peter Morville, President, Semantic Studios. Here is the session overview. My notes follow.

“Search is among the most disruptive innovations of our time. It influences what we buy and where we go. It shapes how we learn and what we believe.  It’s also a radically multidisciplinary, creative challenge. In this talk, Morville defines a pattern language for search and discovery that embraces user psychology and behavior, cross-channel information architecture, multisensory interaction, and emerging technology. He identifies design principles that apply across the categories of web, ecommerce, enterprise, desktop, mobile, social, and real time. He explains how future methods and user experience deliverables can help us to create better search interfaces and applications today and invent the improbable discovery tools of tomorrow.”

Peter said that whenever he comes into organizations he finds unfinished structures for search with no blue prints. Browsing does not scale in these environments. Search is the most disruptive technology of our time. It has major impacts and investing has huge rewards. It demands attention to detail and is more than just putting in a few key words.

Feedback is essential and users need to have confidence in their actions. Design for the context of use (e.g. desktop vs. mobile vs. kiosks).  Real time search should keep giving results as they appear.  Browsing and navigation should align with search but remember that browsing does not scale.

Peter referred to the work of Christopher Alexander, an architect noted for his theories about design.

In Search Patterns Peter looked at behavior patterns such as narrowing. How do you clarify what you are looking for? In some other contexts, broadening is required.  Sometimes people get stuck in their behavior patterns even when it results in a dead end and we need to help people get unstuck.

He also wrote about Design Patterns for search.  He looked at how to help users before getting to results by offering suggestions. Yahoo has auto-complete and auto-suggestion. Sometimes the auto-suggest does not use the original key words but goes beyond them.

Google tries to offer the best first in their design. People often do not go beyond the first few returns. It is also fast, another key to its success. We cannot take slow for search returns.

Flickr looks at interestingness like Google does Page Rank. It looks at social data.

Federated search is a good solution. Other times it is a symptom of a deeper problem.  People do not know what sources to look in. Peter is working with the Library of Congress on this issue. People do not know which database to go to. So they offer a federated search as a band aid for now.

Facted navigation is one of his favorite design patterns.  It allows people to start simple and see the facets or filters that they can use as a next step to clarify.  You can do testing and refinement of the facets and the interface where they are placed. You can also make use of user generated tagging to refine facets.  Amazon is looking at how to apply facted navigation on mobile devices.

Actionable results is another design pattern found in song results. You can do multiple things with the results including sharing and playing.

We can rethink search by asking questions like Wolfram Alpha.  We can try to help people make the right questions like Hunch does.

Peter asked: how can we help people find the unknown unknowns? I like this question, as it is a goal of Darwin.

Ambient Findability is another book from Peter. It covers the degree to which an object can make itself findable. He said the world was not ready for it when he wrote the book in 2005. Now the world is more ripe for the concept.  Some of the ideas are becoming more possible. You can make your iPhone into a bar code reader to find things. Amazon allows you to scan stuff to see if they can sell it to you.

There are new ways to forge connections between physical and digital stuff. He describes a plant that tweets and asks for water when it needs it. His local library has a digital hold shelf for physical books he orders. People are starting to talk about these connections more such as Kuniavsky’s upcoming book on smart objects.

We need to design new maps to better understand the choices people have from a user’s perspective.  Peter likes search because it is wicked cross-disciplinary problem with many inter-dependencies.  It is a problem that requires on-going attention and is never fully solved.

He closed with three ways to move forward: focus on details, see the big picture, and see the problem form different perspectives.  Great session. Here is Peter’s blog.