Re: ::scr Drooling GUI

[celia romaniuk <space@xxxxxxxxxxxxxx>]

Disclaimer: I don't really like Alan Cooper's writing. I'm not sure
why. Perhaps it's because I agree with most of each statement he makes,
but then disagree about the remainder. 

On Thu, 7 Mar 2002, Arvid Gidhagen wrote:
 
> If we take it as a fact (like Alan Cooper does, at least when he's
> talking about the way the file system and documents should behave
> from the users point of view) that the system designers should try
> to construct the system so that it as closely as possible matches
> the user's mental model, I think we'll end up in a lot of trouble
> pretty soon. 


> There is little difference between the statements "the designer should
> try to match the user's mental model of the system" and "user
> interfaces should be intuitively obvious".

First up: one of my pet hates is the word 'intuitive'. 'Familiar' is
much more descriptive. This article explains well:
http://www.asktog.com/papers/raskinintuit.html
(apologies to those who have heard me go on about this before, and had me
death-stare them when they say the forbidden word).

> With a bit of luck, it might work well for a while. But sooner or
> later, something will probably go wrong. If the actual way the
> computer works is too far removed from the "mental model" that is
> presented, a lot of transformation and translation will have to
> be done on the data, and eventually something will happen because
> this translation or transformation isn't 100% accurate, and the
> user will become mighty upset.

My understanding is that it's a two way street. There is learning involved
when using a computer, just as it's involved with use of anything. The
interface, though, can coach and give hints as to what it's doing. 

I think what I'm perhaps trying to get at is that mental models are just
one aspect of an effective, usable system. Every now and then I go back
and read Tog's first principles of design. It's highly relevant to this
discussion. 
http://www.asktog.com/basics/firstPrinciples.html
 
> For instance, it might be nice for the user if there appears to
> be no difference in getting a document from a web server somwhere,
> from another computer on the LAN or from the hard drive. But with
> regards to security and reliability, perhaps it isn't so great
> after all. Web servers can go down, for instance.

See, I'd argue that it's actually pretty important that you distinguish
the locations of documents, precisely because some of the networks will
fail from time to time, and not distinguishing them will lead to confusion
about why the hell the computer isn't doing what it's meant to.
 
Aside: just found some stuff about human frustration with computers, and
other affective stuff, here:
http://arsenal.media.mit.edu/~carsonr/index.html

> I don't think the idea of metaphors in user interfaces is
> completely stupid, though. The danger lies in making them to
> literal, which will cause users to expect that the computer
> actually works just like the source for the metaphor.

Yeah, what tog sez up there above.
 
> By the way, I prefer to be called "cognitive science student", if
> anyone cares. I might end up working with interaction design or
> usability or information architecture when I'm finished, but I'll
> always be cogsci.

:) 

Celia