Re: [tlug] GUI font tools

[Matt Gushee <matt@example.com>]

Stephen J. Turnbull wrote:

>     Matt> Gentoo configs, defoma, fontconfig, etc. ... all the
>     Matt> infrastructure facilities ... will be made to support the
>     Matt> interface. Or else.
> 
> Other way around.

Well, yes, as a practical matter ... tools, physical and otherwise, have 
ways that they want to be used, and I've heard it said more than once 
that good sculptors "discover" shapes in their materials. And my 
suggestion of *forcing* the backends to do the right thing was a bit 
facetious.

But as a design principle, I believe that backends should support user 
interfaces, not the other way around--"the interface is the 
application." There is a tendency in the free software community to 
lavish attention on infrastructure and let the UI take care of itself. 
There's no question that the Bazaar has had some real 
benefits--open-source development tends to produce more robust and 
interoperable software, and that's good for everyone, even if few people 
really appreciate it. But it has failed to deliver much improvement in 
the end-user experience.

I refuse to believe that this is the best we can do. If you read authors 
like Alan Cooper and Jef Raskin (and find them persuasive, as I do), you 
get a sense of how much better computing environments could and should 
be. As for process ... Cooper's recommendations for putting interaction 
design first are very much geared to a corporate setting. Is there a way 
to incorporate the user-centric essence of that approach into a 
democratic/meritocratic development process? I'd like to think so ... so 
far the only way I can see to proceed is individual initiatives by 
people who know and/or care about interaction design (count me as one 
who cares ... maybe I know a little, though probably not enough).

> The GUI should provide enough configurability to
> support all of them, and a plug-in API for back-end shims.

What do you mean by "shim?" I think I vaguely understand, but you seem 
to mean a fairly specific kind of component.

> the majors you don't use just as your "ante".  After that, if your
> interface is a real improvement (by that I mean, as opposed to just
> fitting your style better) you shouldn't have to push to get the ball
> rolling.

Maybe, but it depends on the audience. Hackers like new software, and 
are usually game to try anything out of curiosity or in response to a 
rational argument. But if the meta-objective is "better desktop Linux," 
then you have to be ready to deal with people who are not like that. My 
experience working with non-geeks has forced me to conclude--sadly--that 
many people are terrified of learning new software, even when they 
accept that there are good reasons to do so.

... which could segue into a rant about how so many people these days 
are content with a mediocre spoon-fed reality, and doesn't it suck, etc. 
But I think I'll leave that for another day.

--
Matt Gushee
Englewood, CO, USA