notes from today's class

[Jesse Cirimele]

hicks heiman law:
-measuring the time to pick something given that you know what it is
and it's more complicated than "pre-attentive choice"
-how does this relate to usability?
-airplane pilot: might be a good model to have lots of choices because
log time to add options is possibly way faster than hiding choices (
clicking through screens to find options )

key-action model:
-how many types of keys do we really have? paper says three, we heard
arguments for 2 ( and 4 ).
-taxonomy isn't "right". in order to argue for or against you need to
put it in the context of a task
-argument that descriptive models aren't like predictive models, only
tell you about what kinds of things
-argument that descriptive models aren't generative
-you could use existing descriptions to talk about how you can modify
existing keyboards
-no executive keys no the iphone
-use a descriptive method to generate ideas: flat keyboard with 100
keys and 3 types of keys. what if we changed one of those things?
-being creative is really difficult. it's often by laying down some
rules or constraints you get really creative things.
-also, it's interesting to take a step back and thing about higher
level constraints to allow creative change at lower levels

buxton's paper:
is the three stage model appropriate for modeling a mouse? is added
complexity important?
-depends on the task
-using the models falls apart if your technology can't keep track of
what state it should be in
-models are not enough to predict human behavior. we can't know people
that well.
-one way the model could be bad might be because it doesn't tell us
anything interesting. precision ( repeatability ) doesn't get us
interesting. what does the three state model tell us about ubiquitous
technology? touch screens?
-models are good to explain what is technically capable to give
designers and programmers ideas
-drawing out a diagram to put your ideas into a form that "the
computer understands" is really important in order to translate ideas
into programming. a bridge.

when do you want to put different interactions into one input device,
and when separate them?
-want to have the mappings of the input device to model as closely as
possible the experiences

fits law, mac vs windows
-information theoretic view on how much people can say with a mouse:
about 14bits
-mac one menu on top of screen, windows one menu per window
-if windows wins it's because the distance is farther to the menu, but
in general mac wins. why? two things:
---1) experts want to memorize without looking
---2) target size for mac is "infinite" (don't have to worry about when to stop)

bash predictive text study?
-why just predict instead of testing?

informing design of direct touch table tops
gesture registration - same gesture, different results in different
contexts. is it the right thing to do to overload control like this?
-ipad as small table. making a play system easier than making a work
system. complicated systems (work systems) will need learning
-different physical places might not make sense in the context of a table

more accuracy by using middle finger and thumb to select, finger to select.
-not good. iphone does it fine. don't make the users adapt.
-depends on application: cutting something very small in photoshop is
hard if your finger is in the way, but a game it doesn't matter as
much.

what do we need to do today that is pixel perfect?
-photoshop
-photoshop 1.0 was pixel perfect, now we have algorithms that can do a
lot of the work
-photoshop users, using a stylus, just zoom in to get more accurate
-why do they use a stylus?
-because they are trained as artists they prefer things like a pen,
pencil (experience argument)
-pen gives you fine motor control in fingers vs mouse with elbow,
wrist. fits law tells us that fingers give us 38 bits vs a hand/wrist
of 14bits.
-for tapping or pointing tasks you get a closed loop (see intermediate states)
---better feedback on the stylus (circle where you're going to land)
improves clicking
---if you make the widgets crossing bassed (slash through) then the
stylus performs much better
------does this work with small buttons?
------why would stylus crossing of a small button beat mouse tapping
on a small button?
---------because if you are crossing with a stylus you are dragging
the stylus on the surface and therefor it reduces to the mouse case
but you get the 38 bits of using your fingers