The Future of Search: Peter Morville Defines, Shows Examples
I'm at the Webstock conference in Wellington, New Zealand. Currently speaking is Peter Morville, whose topic is "Ambient Findability and the Future of Search". Peter Morville was co-author of the famous 'polar bear' book about Information Architecture - which was like a bible to many of us who started our careers as webmasters or designers.
The crux of Morville's presentation was that we're "at the crossroads of ubiquitous computing and the Internet", but "the user experience is out of control, and findability is the real story."
Search is still the interface of choice for Web users to navigate our way through all this information and data, but "findability" is the way forward says Morville. Currently, he says, people view the top Google results (say number 1 or 2 on page 1 for a search term) as being the most credible - and as of now this is equated by most people to findability. However there is a lot more to findability to Google search.
Findability in the Real World
Morville defines findability as "the quality of being locatable or navigable". It's becoming more important as information overload takes hold and mobile devices become more popular. As we've written about a lot on this blog, we're in an age where "a wealth of information creates a poverty of attention" (Herbert Simon quote). We also have many more alternate interfaces to information - and this will expand well past smart phones in the future.
So what kind of products will help us achieve findability? Morville says it won't be AI (e.g. a next gen Microsoft Bob) and he also says it won't be visualization technologies (something that perhaps our own Charles Knight might disagree with!). Although he says eventually visualization may play a part. He says business intelligence systems show a way forward - software that find patterns. He also says librarians are important, where "metedata is sexy" - indeed, see Marshall Kirkpatrick's post on sexy librarians!
Morville says it's about bringing together things like tags and taxanomies, not keeping them apart. He points to sites like Etsy, where users can classify their items and tag. He says the future of search will be "a future where search and browsing work together". He refers to a Marcia Bates article about "berrypicking".
Examples of Future Search, Right Now
Morville picks out a number of web products that he says point to the future of search and finadability. One company he mentions is Endeca, a "guided navigation" service which he says mirrors the way our pschology works when searching for information. Buzzillions is another app Morville highlights, which again provides ways for people to narrow and find what they're looking for.
Clusty is a hybrid solution, where humans design the taxonomy but there are automated algorithms to categorize.
Google, says Morville, "is really struggling". However he also says that Universal Search is a good exploration of bringing people into Google's algorithms - as an example a search for 'diabetes' guides people to different types of information.
He also mentions Flickr's clustered tags.
He ends by saying they we "need a lot more innovation". Google Book Search is an example, as is Everyzing which has searchable podcasts and interesting ways to browse it. So this is all expanding the "searchable web" - into audio, video, etc.
He also says that it won't be limited to the digital environment - more and more physical and digital worlds being melded together. aka "Internet of Objects".
Conclusion
In conclusion, he says we need to continue to focus on usability in this new world of findability - "but I hope we don't stop there".
An enlightening talk by Peter Morville and really demonstrated some of the groundbreaking web apps showing the way to the future of search. For more on these topics, subscribe to ReadWriteWeb's network blog AltSearchEngines.
FOLLOW RWW ON TWITTER
RECENT JOBS
Comments
Subscribe to comments for this post OR Subscribe to comments for all ReadWriteWeb posts
I think that current search is brain dead. And I am not too positive in some of the semantic search initiatives. But search can not just be algorithmic and linear, specially since we, humans, do not search for information that way. There are many factors in searching and I am afraid most of the current methods miss them.
Posted by: Fabian Schonholz | February 13, 2008 7:36 PM
Some kind of clustering like Clusty or AllPlus does can be really helpful to searchers, but I think we should figure out how to involve those information more into the search process.
Posted by: Endre Jofoldi | February 14, 2008 12:03 AM
"Ambient Findability and the Future of Search"
Richard, may I ask, if Peter Morville is an expert in internet search? The reason I am asking is he is being invited to talk on a topic that is not his domain of expertise at all. When you get gurus to give talk on topics that in no way related to their field of expertise, it leads to one thing. The guru, in a false sense mislead the audience into thinking that what he says, is absolute truth.
I seriously doubt that Peter has real understanding or deep knowledge about the field of search engine. His comments such as :
- "at the crossroads of ubiquitous computing and the Internet"
- So what kind of products will help us achieve findability? Morville says it won't be AI (e.g. a next gen Microsoft Bob)
- "the quality of being locatable or navigable"
- He says business intelligence systems show a way forward - software that find patterns.
- He says business intelligence systems show a way forward - software that find patterns.
These are all useless babbles. Does he know what AI is? He says it won't be AI, but he also says that business intelligence is the way forward. Does it occur to him that machine learning is AI and also business intelligence is based on machine learning which is AI. He is self-contradicting, because he has no clue to what AI is.
Peter says that Google, "is really struggling". Did he say how? This is an opinionated assertion which has no fact to it. Perhaps he should look up the presentation done by Google guys at the Workshop on Algorithms for Modern Massive Data Sets (WAMMDS) in 2006.
I don't have a problem with self-style gurus doing invited talk about web technology, but I do have a problem when they mislead the audience, especially when they do presentation on topics that are not their domain of expertise and this includes Peter.
Perhaps, the next time Webstock, is inviting experts in the domain of search, I can help the organisers with some names of who is who (researchers) in the field of search technology. Definitely, any one of those authors who did presentation at WAMMDS (2006) is an expert on search technology.
Posted by: Falafulu Fisi | February 14, 2008 2:37 AM
For those of you seeking to experiment Buzillions-like refinements on the whole web scale, you can try Exalead.com :
available refinements include automatically generated n-grams, DMOZ categories, file types or media types. Plus, the Wikipedia search http://www.exalead.fr/wikipedia offers semantic refinements about people, locations and organization related to your search.
Nicolas Steegmann
Exalead web product manager
Posted by: Nicolas Steegmann | February 14, 2008 5:26 AM
The following article is an excellent one to read about the direction of search and its associated technologies & algorithms. The 10 page article was written by a young Professor Amy Langville. Amy is one of the world top expert & researcher in search algorithm & data-mining. She came up with a very fast algorithm of how to calculate Google PageRank algorithm in 2003 (the power method of how to solve PageRank, there are other methods of how to solve PageRank). I believe that the folks at Google adopted Amy's method which dramatically improves the PageRank. Amy's work was publicly available , ie, it was published in a peer review journal, so Google guys are scouring such journals non-stop to look for interesting ideas. The article was written to target the general public (ie, no equations involved), so it is readable to anyone. I have had a few email exchanges with Prof. Langville in the past, as I had used some of her matrix algebra codes (available from her site) and also I frequently ask questions to clarify the algorithms in some of her papers (most are freely downloadable from her site).
The Linear Algebra Behind Search
An unrelated topic but relevant. Does anyone from the US know why professorship is awarded to young PhD academics? I assume that Prof. Langville is in her late 20s or early 30s (based on chronology of her academic study that is described in her website). Professorship here in New Zealand is awarded mainly when an academic are in their 40s or above. I don't doubt Amy's professorship, since she is one of the leading expert on search algorithm, but it is so common that I see young professors from the US at similar age to Amy.
I have a suggestion Richard. Perhaps if RWW or AltSearchEngine could invite Prof. Amy to write a featured article for your site. I am pretty much sure that she will be keen to do that.
Posted by: Falafulu Fisi | February 14, 2008 11:16 AM
The bottom line with search is that the query interface is not broken by any means. It's ubiquitous across the internet, and, it works. It's the search results that need improvement.
As long as you are using a web browser to do a search then you need to keep the qty of clicks to a minimum. We still have a stateless framework, regardless of how much AJAX seems to blurr the lines.
Posted by: Marc | February 14, 2008 1:19 PM
Interesting!
I agree with this article.
I think usability is very important and it will become increasingly important.
Posted by: Mariachiara Marsella | February 15, 2008 2:39 AM