I find the study of software user interface fascinating. Not so much in the area of efficiency, guidance, or discoverability; I'm fascinated by what developers identify as "good" UI. With few exceptions, developers fall into some variation of this fallacy:
The whole idea of good UI is to provide a good experience for someone else, to make someone else more efficient, to reduce the user's level of anxiety. But developers turn this into a self-centered reversal of common sense, assuming that they know what works as well as any user of their software.
If these assumptions had a basis in cognitive psychology, color theory, or extensive testing, I'd more willingly entertain them. But very often, good or bad boils down to "I like the way that looks." Good and bad UI gets boiled down to a static impression of color and widget spacing. The dynamic aspect of the software is largely neglected, in favor of solutions like UI skins or themes.
Why am I thinking about this? Some weeks ago, Sun released a video of Project Looking Glass, a technology demonstration exploring what might be done with three-dimensional, hardware accelerated interfaces. It is very clearly a technology demonstration, intended to demonstrate possibilities, to get people thinking about different interface metaphors, ones where elements are directly manipulated in a way that more closely resembles the physics of real-life.
But the reaction generated by Sun's research project has included more than its share of this kind of stuff:
Setting aside the fact that Project Looking Glass is not intended to ship, but to generate ideas, Buzz is missing something. Yes, the type-ahead live searching in iTunes is excellent for finding a specific song or album, but it is not well suited to browsing. A music collection is just that: a collection. Part of the enjoyment of having 30 Gig of music on my laptop is in picking the album. Almost invariably, I browse through lots and lots of albums. Sun's demonstration suggests a somewhat different way of browsing through a CD collection. Buzz assumes it is better to zero in, like a missile, on a specific CD. [This isn't a specific criticism of Buzz Andersen. He is a smart dude who has written some greatstuff. His 'blog just got me thinking.]
The iTunes type-ahead feature is good for searching, but totally inappropriate for browsing.
This is the kind of distinction that becomes absolutely obvious during real user testing, but might be entirely missed if left up the the anecdotal experience of a handful of developers. In fact, some developers have noticed this problem, trying to solve it with clutter. But clutter is still a two-dimensional, largely static tool. Would it be more effective in three dimensions?
Take-away message: user interface isn't about what I like; it is about what works.
Good UI is UI that I like.
The whole idea of good UI is to provide a good experience for someone else, to make someone else more efficient, to reduce the user's level of anxiety. But developers turn this into a self-centered reversal of common sense, assuming that they know what works as well as any user of their software.
If these assumptions had a basis in cognitive psychology, color theory, or extensive testing, I'd more willingly entertain them. But very often, good or bad boils down to "I like the way that looks." Good and bad UI gets boiled down to a static impression of color and widget spacing. The dynamic aspect of the software is largely neglected, in favor of solutions like UI skins or themes.
Why am I thinking about this? Some weeks ago, Sun released a video of Project Looking Glass, a technology demonstration exploring what might be done with three-dimensional, hardware accelerated interfaces. It is very clearly a technology demonstration, intended to demonstrate possibilities, to get people thinking about different interface metaphors, ones where elements are directly manipulated in a way that more closely resembles the physics of real-life.
But the reaction generated by Sun's research project has included more than its share of this kind of stuff:
"My answer? Well, judging by the provided screenshot, it would take a lot longer to find what I’m looking for in my music collection since I would be scanning through hundreds or thousands of literal images of CDs instead of performing a lightning-fast type-ahead text search in iTunes!" [via SciFi HiFi]
Setting aside the fact that Project Looking Glass is not intended to ship, but to generate ideas, Buzz is missing something. Yes, the type-ahead live searching in iTunes is excellent for finding a specific song or album, but it is not well suited to browsing. A music collection is just that: a collection. Part of the enjoyment of having 30 Gig of music on my laptop is in picking the album. Almost invariably, I browse through lots and lots of albums. Sun's demonstration suggests a somewhat different way of browsing through a CD collection. Buzz assumes it is better to zero in, like a missile, on a specific CD. [This isn't a specific criticism of Buzz Andersen. He is a smart dude who has written some greatstuff. His 'blog just got me thinking.]
The iTunes type-ahead feature is good for searching, but totally inappropriate for browsing.
This is the kind of distinction that becomes absolutely obvious during real user testing, but might be entirely missed if left up the the anecdotal experience of a handful of developers. In fact, some developers have noticed this problem, trying to solve it with clutter. But clutter is still a two-dimensional, largely static tool. Would it be more effective in three dimensions?
Take-away message: user interface isn't about what I like; it is about what works.
posted by Eric Hancock at
2:26 PM
0 Comments:
Post a Comment
<< Home