QVGA V/s VGA for GTA03 (was something about yummy CPU-GPU combos!)
Carsten Haitzler (The Rasterman)
raster at openmoko.org
Mon Jun 9 12:01:29 CEST 2008
On Mon, 9 Jun 2008 08:56:22 +0800 Carsten Haitzler (The Rasterman)
<raster at openmoko.org> babbled:
> On Sun, 8 Jun 2008 18:58:15 +0200 "Dr. H. Nikolaus Schaller"
> <hns at computer.org> babbled:
> > > we could just not ever even ask you guys and you get what you are
> > > given. too
> > > bad. no input at all. i've opened up the floor for input - but i'm
> > > trying to
> > > dig specific things out of it - not things that smell of"i just want
> > > higher
> > > specs". or keeping up with the joneses. i want real use case
> > > scenarios that
> > > make real sense. :)
> > This discussion starts to become quite boring. Isn't a single
> > potential customer who says
> > "I want it and I am willing to pay for it" enough? There have been
> > several here on this list,
> > if I remember correctly who expressed exactly that.
> no. it is absolutely not enough. why? i am asked by product management to do
> things that are just not possible in vga (to do sanely/fast). they come first.
> you users come second. in the end if product management want X they get X. and
> if for X to happen we go QVGA, then so be it. you guys lose. i need a very
> very very strong argument against going to qvga - and that means product
> management need to drop a feature.
note - i am talking hypothetically. i don't want to discuss vga as a product
management feature - not if you like it or not, or it looks pretty. i am
looking for hard cold technical facts. what does it stop being possible
1. u may need to scroll more
2. viewing of images/data that just have more pixel content will need to be
zoomed out and have less display fidelity
3. some things requiring text displays like 80x24 terminals will be not
readable at all at font sizes able to fit on the display (they will jut blur
away all character details).
with almost everything i can think of you can get by qvga by:
1. scaling data
2. changing font sizes
3. re-arranging ui elements etc.
no matter what you need to do this even for vga - if coming fro xga land or
better. it's just a more extreme case.
no mater what at vga - u still need to zoom most web pages. even at 800x480 you
still need to. i have a n800. i know how often i have to scroll horizontally
even with 800pixels to play with. i know what it ends up looking like. so qvga
is just a more extreme level of zooming or scrolling needed.
an 80x24 terminal is possible to make it readable @ qvga - if we allow
scrolling. (and possible in landscape with an ultra-tiny 4-pixel wide font -
possible (3 pixels for text, 1 for space). not very nice though.
at some point someone will have to make a call on resolutions. maybe we make a
much smaller phone with a smaller screen and thus you will need to have fewer
pixels anyway? who knows. but if there are uses that cannot be somehow crammed
into qvga, i would like to know.
right now freerunner is vga - and nothing will change.
gta03 is also vga - it is theoretically possible to change without much impact,
but chances of a change are very slim, unless qvga is a "that's fine for
everyone" ANd product management want to push it. right now they don't push one
way or another.
as for future phones - who knows. but knowing what you guys do, want to do, and
need is important. so we need to think of more virtual framebuffer technology?
(eg advertise a higher res but scale down with a compositor?). is high-res an
absolute must for functionality for your uses, or just a "nice to have" to show
> > BTW: a use case doesn't say anything about required quality. It
> > describes a sequence of interactions
> > (http://en.wikipedia.org/wiki/Use_case). Sorry, but I can't disclaim
> > my academic history :)
> i asked for use case because i am not just talking quality. i am talking a
> case where vga makes something possible at all or not. where something just
> wouldn't be usable or possible without vga. that is what i asked. i want a
> use case for vga. not just a "it looks a bit nicer".
> > Nikolaus
> > _______________________________________________
> > Openmoko community mailing list
> > community at lists.openmoko.org
> > http://lists.openmoko.org/mailman/listinfo/community
> Carsten Haitzler (The Rasterman) <raster at openmoko.org>
> Openmoko community mailing list
> community at lists.openmoko.org
Carsten Haitzler (The Rasterman) <raster at openmoko.org>
More information about the community