Something that I got from Yahoo presentations at WWW was their interest in moving things in the UI space. Google has famously held fast to the simple 1-d search UI, and I’ve argued that this is very important to them as it keeps focus on the power of their relevance algorithm. If they lost control of the UI, then approaches to search that relied on more emergent behaviours between the user and the data could appear, changing the game.
I think there are 3 forces that determine the nature of a UI. In no particular order, they are the user, the developer and the data. Understanding what each brings is important when designing applications which involve large volumes of data, and in strategic positioning and paradigm formation.
The user’s influence in the UI can probably best be seen in Google. Here, Google celebrates a metrics driven approach to UI evolution and goes where ever the user takes them. They have to conform the laziness of the user – who doesn’t want to be challenges or to learn new skills before getting back value.
The developer’s influence is obvious – the developer creates the darn thing, but they also have to manage and support the code base.
Finally, the data’s influence on the UI refers to how closely the presentation fits the (representation of the) data. If you have geographic data, the UI is a map with pins, if you have temporal data, the UI is a time series. What I find interesting here is that the closer you get to showing the true nature of the data, the more opportunity there is for errors to emerge. If I place a pin on a map (a very precise representation) I better have that pin in the right place. If, on the other hand, I provide a general indication of the location of the content and then hedge by presenting the data as text, the system can avoid problems of accuracy.
>> I think there are 3 forces that determine the nature of a UI.
>> In no particular order, they are the user, the developer and the data.
Bzzt. Wrong answer. Replace "developer" with "UI designer" to get the right answer.
Posted by: Dmitriy | May 02, 2009 at 02:22 PM
"If, on the other hand, I provide a general indication of the location of the content and then hedge by presenting the data as text, the system can avoid problems of accuracy."
Is there a Data/Search uncertainty principle?
Posted by: matt | May 06, 2009 at 06:26 PM