::scr Ramblings of a Classic Refugee or How I Learned To Stop Worrying and Love OS X

[celia romaniuk]

On Sun, 3 Feb 2002, simon wistow wrote:

> I was tangentially involved in a project that was working on genetic
> interfaces. Which worked like this ...
>
> The user plays round with an application. Depending on which items of
> functionality (hence forth to be referred to as functions) they use the
> genes of that function get stronger and breed more. Essentially. There's
> more to it than that but it gets a bit involved and then I go crossed
> eyed.

The nice thing about this type of development is that you get people
working with an interface on a fairly constant basis - as opposed to
having sessions where you work with someone and observe and/or ask them
about what they're doing (the kinds of techniques described at
http://usability.gov/methods/same_way.html

On the other hand, the great thing about doing it in person is that you
can ask questions as you go along, and observe, and implement design
tweaks to problems that the user may not be able to articulate or
recommend a solution for.

The classic example (from the web - OK, so yeah, I'm a webmonkey) is of
that of people using the search function of a site. That may mean 'make it
stronger'. But it may also mean that they can't find what they want by
browsing the pages. Having said that, maybe people use it cause your
search function rocks c.f. amazon. Thing is, you're only going to know why
people are using it if you ask them.

> After a while you regenerate the interface automagically from the
> genetic pool. Thus the more useful functions become more prominent.
>
> There are three things to think about:
>
> 1. Do you do this during testing and then release a finalised version of
> the interface hard coded in?
>
>    1.1 Can you then produce three versions of the code for beginners,
>        middle and power users. Or people who use, say, Word, for writing
>        letters or for producing proposals. Or have the interface swap
>        over depending on what mode you want.

Yeah, that's an interesting one. It did occur to me that people who were
beginners may use some functions more than others, and this would skew the
results. As you point out, Raskin does so 'no no designing for different
levels of users BAD', but then again, Eazel's Nautilus (which looked very
cool, and involved ex-Apple people) was going to utilise it.
http://web.archive.org/web/20001018180419/nautilus.eazel.com/screenshots/aug-02-2000/index.html

Personally, I'm undecided on this one. But my gut-feeling is that if the
basic 'beginner' controls are obvious, and it's clear how to use the more
advanced features, then people will help themselves. Much as skilled
Photoshop users love that app, it's something I hate about it. I dabble in
it occaisonally, so know how to do basic stuff (change image size and
colour mode, crop, blah blah blah) but so often the way to do things is
unclear unless you have someone over your shoulder telling you. For
example, I can never remember how to apply to apply a filter to a channel.
And there's no 'step by step' guide for people like me, who can't
remember.

>    1.2 Or do you have the interface automatically update? Tailoring
>        itself precisely to the user. This is effectively what Microsoft
>        have done with their latest versions of their interface and
>        everybody hates it but there are more sophisticated, less clumsy
>        ways to do this. That way you don't need to go back and fix the
>        interface, you just give it a good start and it iteratively
>        makes itself better.

Hm. Yeah. I agree about the MS clumsiness - it's awful. But perhaps by
assuming that people fall into expert-bgeinner modes is presumptuous.
Given my example above, how would you class me? Because I'm a beginnerish
user but I want to do expert level things. I think that's the crux of the
argument.

> 2. Could you have the users machines constantly sending back information
> that will let you build the next generation version of the application

See above, about actually talking to people (gasp).

> 3. Something [forgotten], I'll come back to this later. Probably.
>
> (later ... nah, still can't remember it, one day I might remember it)

Still blank?

> Of course Raskin would probably scream about this. Especially the
> 'different modes' bit and certainly the changing interface from
> underneath you. I think that there may be a middle ground soemwhere, I
> just can't put my finger on it.

Yeah, he would. But he's just one guy, you know.

On a tangential note, I saw this today:
http://www.edge.org/q2002/q_hurst.html
which is  Mark Hurst from Creative Good (
http://www.creativegood.com/ ) talking about why these fancy new
interfaces are a bad idea. Just thought some of you would be interested.
Personally, I think that there's a strong middle-ground - make interfaces
better, and help people to understand them. But then I'm a soft wooly
cardigan-wearing liberal.

And I still think that the BeOS way sounds brillient. (Ss posted by Paul
earlier, but here it is again, cause it's that good, honest:
http://www.birdhouse.org/macos/beos_osx/ )

and by the way again, there's been a lot of talk in Information
Architecture-type forums lately about facets, and I wonder if they've seen
it. I should really post to one of their lists one day, rather than
hanging round in brane-exploding scary-l33t forums where I don't belong :)

-- 
celia
seesmm quite peacuful to me