RGB pixel direction question Re: Drawing raw pixels to screen..
richard.franks at drdc-rddc.gc.ca
Thu Nov 23 21:05:10 CET 2006
> Salve Richard!
Salve Rob! (? - Being British makes using foreign languages scary)
> > creative(ly different) UIs or branding constraints. That is - your
> > standard phone interface is limited by the *perception* of what the
> > lowest common end-user denominator can handle.
> Does it must be so fixed or couldn't be it more flexible that it
> has two or three different concepts?
Moreover that flexibility, has historically meant complexity - Linux and
Windows demonstrate different yet roughly proportional quantities of
both. I don't see why a framework can't be designed to maximise
flexibility and minimise complexity -- but it's near impossible to grant
that to a system which is already established.
In terms of the established companies designing UIs for phones/PDAs, the
risk of shipping a device with an unconventional interface is high -
which may be why we see such parity between different models.
> > Fundamental to this though, is how quickly the Neo1973 can update its
> > display - is it possible to get some FPS figures for just a simple for
> > XY loop?
> seeing a zune presentation, but also some new smartphone
> - I don't see a benefit of TO MUCH animation in the GUI
> - only because it is possible or new....
Oh gosh - yup - agreed. However, intuitive designs such as the iPod
scroll-wheel show that stepping out of the archaic Xerox
menu/icon/pointer paradigm can: a) hold mass-market appeal, and b)
address specific constraints with regards navigating masses of
I think there's still a lot of room to explore, because we have no
share-holders to mollify!
That said, my embedded experience is of the roll-your-own cross-compiled
kernel, and write your own application software variety, I haven't
dipped my toes into the mobile world yet.. so my market-assumptions may
be completely off-base?
> So can you explain your fear that the update of the display would
> be to slow?
a) I'm developing a 3d engine from the bare-bones, which should (!) be
efficient enough to run on embedded platforms such as the Neo1973. If
drawing-to-screen is overly expensive though, there's no chance.
b) A good UI (IMHO) should blur the distinction between content and
'fancy effects'. In that access to content is prime, and 'fancy
effects' (compare iPod scrollwheel vs Vista window swapping) should
complement access to content rather than obfuscate it.
More information about the community