Saturday, November 26, 2005

Interfaces, Part III

Why is it so hard to design good interfaces? And what is it that makes the difference?

With no training in HCI, I speculate that it has to do with how closely the interface correlates with the world we're used to interacting with. I have no idea if this is true, but it seems reasonable, and so it's going to be my premise for the moment.

(An aside: if this is the case, there are two ways to handle it: change interfaces to be more like how we think, or change how we think to be more like some kind of interface. The latter is hard to justify, since, if nothing else, how do we pick what we bend our minds to? Not that some more education of the general populace would be bad, but this seems like a homogenizing effect, which I'd call unabashedly bad.)

So, assuming this is the case, why should it be hard to make something which is like the real world? Partially, I think, because it is a matter of consensus. It's not just my view of the world, it's figuring out what the consensus view is. But not just the consensus, because there mostly isn't one. It's like factoring out the common elements that most people have. It's a level of abstraction of figuring out what interacting with the world means without actually interacting with the world. And it's worse, because you have to figure out the common elements in this level of abstraction for a mass population.

Mass populations can be very hard to predict. On some levels, they're just statistical, and this works pretty well. On others, they're the hardest thing to predict, since really, you have to predict each element, and each element is extremely complicated in itself. Since we still suck at user interface, I'm going to guess that this isn't one of the things you can just do statistically (or if it is, we haven't found the right parameters). Thus, interface turns into a ridiculously complicated probem if you want to do it right.

(Actually, really doing it right would probably mean an adaptive interface that would reconfigure on the fly to suit each user indivitually.... but that is another matter entirely.)

I got to thinking about all this when an aunt sent me this article:

http://www.edwardtufte.com/tufte/vitalstats

It's not directly related, but data presentation is just the problem of interface. What's your interface to data? Is it really that different an issue whether you're looking at email or correlated statistics or a bus schedule? You're still trying to put it in some framework that somehow "fits" into a structure of the brain to trigger the right understanding.

One line I must point out that I really like: "Same picture, but many stories" (Tufte, from the article). This brings out another aspect of interface to consider: conveying multiple meanings at once. This might be a good thing or a bad thing depending on what you're going for. I think more kinds of data might lend themselves to this than is immediately apparent. Who would have thought a bus schedule would lend itself well to multiple interpretations or evocations? (On the other hand, if you're just trying to find the next time at your stop, maybe you don't want to get distracted looking for the aerial shot of your house.) What other kind of data is really boring, stuff you'd think should just be presented unambiguously? How about finanicial data? Let's take this one and run with it for a moment. What if you take some nice boring accounting stats, and present them in a way that brings out some obscure trends (or at least, where you can see different trends by looking at it in different ways)? Further, do it in a way that it activates different parts of the brain and gets people thinking creatively. Maybe this could get someone going on the obscure trends (since they're feeling creative) and come up with something useful that they wouldn't have otherwise?

Another point I take away from that article is that this is one place science meets art (not that it shouldn't actually in a lot of places... but that's another matter too). At its best, the presentation of data should inspire new connections, tell new stories, and evoke new thoughts and feelings. And here we are, as often as not, using it to bend the truth.

But why should you listen to me? I'm just some weirdo sitting at a keyboard with white goo on her face... who, on top of this, managed to slam a door on her finger yesterday. Stupid finger got what was coming to it for not getting out of the way.

0 Comments:

Post a Comment

Links to this post:

Create a Link

<< Home