XO-2 software plans

John Gilmore gnu at toad.com
Thu May 22 18:17:02 EDT 2008

> What are the software plans for the second-generation XO?

First they need to build one out of something other than modeling clay
and Photoshop.

Then whenever your hand comes close to the laptop, ugly black bars are
going to cover all the edges of that nice sky-blue screen.

There's no need to go crazy, it'll evolve from the current software.
Since it's 2010, assume higher integration: many fewer chips, more
MIPS, lower power, more RAM, more flash.  Brighter, higher resolution,
cheaper, lower power screens from Pixel Qi.  A system-on-chip designed
to suspend in milliseconds at very low power.  Linux.  Since the high
volume very low cost market will have proved itself, you'll be able to
get actual documentation for the hardware.  (Intel will decline to
participate, for that reason.)

If they'd put me in charge, I'd make sure it wasn't an x86 CPU, so
this pesky Windows nonsense wouldn't come up.  Nobody argues that a
non-x86 has to run "Windows CE to prepare students for business
cellphone use".  If you've interacted with a few farm animals, you
don't have to be told where not to stand when encountering a new one.
The same goes for operating systems.  The basic lessons are: There's
usually a way.  It's always harder than you want it to be.  Be sure
your backups do actually restore.  Listen when other users gather.
Beyond that it's just details, and you can pick them up quickly.

You'll be able to suspend for low power without running the Sugar GUI.
You'll be able to share applications without running the Sugar GUI.
You'll be able to mesh wifi without running the Sugar GUI.  You might
even have a GigE jack to reliably network a school full of computers.
All those capabilities will migrate into the infrastructure, where
they belong.  Once there's a choice of GUIs, Sugar as a "look" will be
as popular as the IBM 3270.

Somebody, probably not at OLPC, will finally write a decent book
reader for it -- in 2023 or so.

It'll be hard for OLPC to get multi-touch working when for the last 15
months they haven't had the bandwidth to figure out whether the
current touchpad can do "tap to click" (ticket #959).  But developers
and users of devices built between now and then will write most of the
software needed.  The free software ecosystem will save the day again.

	John  :-)

More information about the Devel mailing list