Search is in the news a lot at the moment, as opposed to just serving it up. You’d think it would a done-and-dusted market, with Google sat on top of the indexing mountain, but no. There’s a whole unexplored world of opportunity in this particular playground. Some articles that cropped up over the past week:
A head for detail
Article on FastCompany describing the MS Research project MyLifeBits and how Gordon Bell is recording every part of his day using a digital camera and audio recorder. It’s a great article that explores the potential positives and negatives of having our lives recorded. I was about to say memories instead of lives, but that would be wrong – our memories do not necessarily match what actually happened… and that would be where some of the negatives are highlighted. We tend to view the world in various tints – rosy when it comes to what we do individually, sometimes not so rosy when judging others. A camera doesn’t record in tints (not memory tints at least). There’s lots of cognitive research on this subject, and great books available. One that focuses specifically on this issue is ‘A Mind of Its Own‘ by Cordelia Fine, a great little book that talks through why our brains deceive us and why that is a good thing… On the positive side, recording memories has the potential to alter how we work, freeing up our brains to do other activities, potentially becoming more productive and creative. And recording daily activities could potentially help with memory-related health issues, as well as the simple need to recall faces and names… But the really interesting part of the article describes what researchers have done with all the information being recorded by MyLifeBits. Facetmap is a MS Research prototype that shows how Bell’s information is connected by using principles from cognitive science – we organise our memories by time and people and associated events (who can’t remember where they where and what they were doing on September 11th… don’t even need to give the year). An interesting application has already been discovered. Alan Smeaton, at Dublin University, set up some students with SenseCams. At the end of each day, they would spend a minute just scrolling through the thousands of pictures taken. The result? An improved short-term memory. As described by Smeaton in the article:
“You actually remember things you’d already forgotten. You’d see somebody you met in a corridor and had a two-minute conversation with that you’d completely forgotten about. And you’ go, ‘Oh, I forgot to send an email to that guy!’ It’s bizarre. It improves your recall by 100%”.
Interesting stuff, and lots of other ideas covered in the article, including the difficulty of how to make a computer ‘see’ the contents of a photo. A hot topic of research at various companies right now and a solution to which would certainly benefit the likes of Flickr, YouTube, MySpace, GoogleEarth etc.
The race to create a ‘smart’ Google
Article in Fortune Magazine looking at Recommendation systems and their influence on our purchasing decisions as well as their ability to improve search results by understanding the context around our questions. This is a tricky area, with lots of potential benefits and equally as many potential pitfalls too. The article talks through various companies trying to build the perfect algorithm for generating recommendations and building relationships. An interesting snippet towards the end of the article – Google’s director of research, Peter Norvig, is an adviser to CleverSet, a recommender company in Seattle. A quote from Steve Johnson, CEO of ChoiceStream:
“Google needs to go to a preference-based search paradigm.”
Beneath the metadata – David Weinberger
And finally, stepping back from research a little bit. An interesting debate sparked off last week on the pros and cons of taxonomy versus folksonomy. David Weinberger’s post captures the main talking points. I even left a comment there myself. Its a debate that is unlikely to be settled soon and I have to confess I am more on the side of the taggers than the taxonomists. I’ve watched a fair few companies struggle with creating and implementing taxonomies over the years, the rigidness of the structure rarely fitting well with everyday activities. I do think a limited taxonomy can be very beneficial to an organisation, but take it too far and it usually makes the right information harder to find when it is supposed to become easier. Tagging has enjoyed great success in the consumer world, as demonstrated by Technorati, Del.icio.us and Flickr to name the most obvious few. It’s time organisations were able to experiment with it too.
Without doubt, in my opinion, the biggest missing feature from Microsoft’s SharePoint 2007 is the ability to tag content, and then view content based on tags. And it’s a real shame because every customer I’ve worked with over the past six months would benefit from tagging, and would use it. In the meantime, they are struggling with how to implement a taxonomy instead… Frustratingly, the first version of SharePoint (SharePoint Portal Server 2001) did include tagging, albeit a limited feature that didn’t work as well as planned. The feature barely made the cut for the next version and has now been removed completely. I hope its removal indicates a possible replacement coming out of MS Research.
If you want to keep tabs on what Microsoft is up to around search, Mike Pallot – UK-based product manager for search and high-performance computing – is blogging.