[IxDA Discuss] Fitt's law for mobile devices

[Peter Bagnall]

On 14 Aug 2006, at 16:30, Barbara Ballard wrote:
> I've been working on articulating some of the theory behind mobile UI
> design, by examining current algorithms and heuristics from the entire
> field.  I'd like some feedback (blog post at [1]).  I've got further
> analysis with specific design recommendations at [2], but that is a
> topic for a different message.


I know you asked for feedback on [1] but it's actually [2] that I had  
issues with! [1] seems ok except I think your concept of "close" is a  
bit confusing in the context of Fitt's law. Number of buttons presses  
cannot be used as a measure of distance for Fitt's law.

Fitt's law was originally formulated as a way of predicting the time  
taken to move of people's hands moving towards a target. It's been  
used successfully to look at mouse movement, but that is an extension  
to Fitt's law that has been validated and is still dealing with limb  
movement and coordination.


in [2] you say...

"Stylus driven interfaces (small screens) – the concept of “distance”  
is almost meaningless...."

I see what you're getting at, that the distance, being small, is less  
important than the size of the target, and that's entirely correct,  
but I'm a little uncomfortable with the "almost meaningless". If the  
target is very small the distance will become important again. This  
after all, is exactly what Fitt's law was originally all about. It  
talks about the distance/target size ratio as being the determining  
factor in time to target.

"Scroll-and-select interfaces (small screens) – the number of  
keypresses to access a target is a good measure of distance."

Again, I see what you're saying, but in this case this is not Fitt's  
law. There is no coordination to be done. If the thumb or finger is  
already on the button, distance is zero, and Fitt's law is not  
helpful. Instead it's more about cognitive issues. I would suggest  
something like GOMS would be a more appropriate way of looking at  
this, although I could be wrong there, it's not something I've ever  
really used myself.

For Fitt's law distance is physical distance - as soon as you stop  
talking about that you are no longer talking about Fitt's law. To me  
it seems that you're pushing Fitt's law beyond it's validity. That  
doesn't mean what you're saying is wrong, just that Fitt's law is not  
supporting your case. I actually think your conclusions are correct!

Apologies if this seems overly nit-picking, but one reason many  
people think that "studies" are dodgy isn't because the study itself  
is poor, but due to over-generalisation of the results. Most studies  
are pretty specific, and you have to be very careful when you  
generalise to be sure your claims are still valid. Most of what is  
said about Fitt's law in an HCI context is actually a number of  
implications of Fitt's law in specific circumstances, many of which  
have been validated experimentally in their own right. For example,  
there's nothing in Fitt's law that says it will apply to mouse  
movement. That has to be tested separately to establish it's validity  
(and I'm pretty sure it has been although I don't have a reference to  
hand).

Cheers
--Pete

----------------------------------------------------------
As long as people believe in absurdities they will continue
to commit atrocities.
		Francois Marie Arouet (Voltaire), 1694 - 1778