[Live-demo] Teaching Spatial Programming - Raspberry Pi

[Hamish]

[housekeeping: the mailing list software gets upset if you put too many
people in the header, if you're on the live-demo ML feel free to cull down
the cc: list a little; also to help dealing with the high volume of email
please cut down the >>>> quoted chain to just the relevant text you are
responding to, including who wrote it, so humans can parse it faster;
pls avoid top posting; etc.. (basically just the usual Usenet etiquette)
tx]


Alex wrote:
> Yes, it's a different cpu architecture and could require some
> changes to ensure compilation. Debian maintains an ARM build but I
> think QGIS or GRASS fails to build on that arch.

GRASS builds fine on ARM (it's ANSI C / C89 so pretty much lowest common
denominator dependency-wise & should build on any hardware or OS from the
last 20+ years). Debian's QA system reports no errors building QGIS 1.7.4
on the armel platform:

http://packages.qa.debian.org/g/grass.html
http://packages.debian.org/search?keywords=grass
http://grass.osgeo.org/images/zaurus_grass.jpg
http://grass.osgeo.org/images/ipaq_grass.jpg

http://packages.qa.debian.org/q/qgis.html
http://packages.debian.org/search?keywords=qgis

And the new "armhf" (ArmHardFloatPort) ARM7+ port seems to be going well
too.



Ubuntu's armel port seems a bit thin on packages by comparison, but I
think that will be Ubuntu HQ's issue and not the specific packages.

http://packages.ubuntu.com/search?arch=armel&keywords=<package name here>
 (currently just seems to be platform="all" doc & data pkgs getting through)


> As far as osgeolive goes it would be a special VM or emulator of an arm
> chip or an ARM based machine that we'd do a special ARM build on.

(i.e. Qemu+KVM: actually Google's Android SDK VM is really just a
stripped down, repackaged, and tuned Qemu bundle)


> Long term yes, I think a RasberryPi could be used for teaching
> geospatial.

Seems to me better suited for field datalogger/data entry use, rather than
desktop GIS where you might want 2 monitors on a grunty workstation.
As always, it depends on the specific task I guess.


> OLPC would probably happen sooner since that x86 based
> (could likely happen now).

I'm pretty sure it already has. I've got a vague recollection of someone
showing me QGIS/GRASS running on one. Valuable if for no other reason
than the daylight readable Pixel Qi screen. The current crop of consumer
laptops all have the high-contrast-arms-race reflective screens which are
unusable out in the field on a sunny day, even with cardboard, black
garbage bag, and duct tape visors assembled. The only big players still
shipping matte finish displays seem to be ToughBook, but they are ~6x as
expensive. :-(   but I digress.. :-)


best,
Hamish