Re: Focus of the project

[I'm not trying to make any claim to authority by responding to this,
but here are my thoughts.]

On Tue, 2002-07-02 at 22:11, George Caswell wrote:
> I am considering switching from my current army of Palms
> to a Linux handheld, but unfortunately it seems that most Linux environments
> for handhelds are pretty much just clones of WinCE.

From my brief experience of WinCE I didn't notice much worth copying, so
that's certainly not what I'm trying to do by my contributions to GPE.

> I think the PalmOS
> approach to user interface and certain aspects of data management is actually
> quite well-done and very comfortable in practice. It makes very good use of
> screen space in the built-in apps, common operations are quick, and the system
> as a whole is very responsive.

These are all good aims for any PDA-style system. I'd argue they're also
aims that GPE has already been following. We're coding in C, which makes
the system as responsive and quick as possible while maintaining
portability. We're using a window manager (Matchbox) which allows the
user to make the most of the screen space. We're intending that the core
applications should share data, while themselves remaining cleanly
separated. We're starting from a solid GUI platform (GTK+ on X), but
have a GPE widget library to override interfaces which would be
inappropriate on a handheld device.

> I'm hoping that if I upgrade to a device with
> a high-res color screen and more horsepower that I can find a way to retain
> those strengths. I rather suspect that, as simple as a PalmOS-style system
> seems, implementing one on top of Linux may be more than I could handle on my
> own.

Well, if you get involved with GPE at this stage you'll have a good
chance of influencing it to become a system which you'd find useful. :)

I'm certainly very aware of the importance of good human-computer
interfaces, especially on PDA-style devices. You don't want to have to
think about the interface you're using, and you don't want it to waste
your time. Given the kind of system we're trying to create, I think we
should be careful to try out different interface design choices, so that
we can experiment to find good ones, and not just stick with what we
think of first - we should be ready, without embarrassment, to throw out
interfaces we've already implemented if we find deficiencies in them or
think of something that people find better.

PalmOS's example embodies a straightforward principle: you want the most
common tasks to be achievable as simply as possible, and you don't want
the interfaces to achieve rare tasks to get in the way. That's something
that any user interface should be trying to follow, but often people
forget about it and try instead to show off all the system can do....

-- 
Moray