[IAEP] [beagleboard] Why Embedded Sugar?

[David Farning]

On Tue, 2008-08-19 at 21:29 -0500, Bill Gatliff wrote:
> David Farning wrote:
> > It has been a bit of a plug but it looks like we have reach critical
> > mass for a self sustaining embedded Sugar community.
> 
> I love the idea of getting a critical mass around something, but I don't yet
> "get it" regarding Sugar for embedded work.

<snipped>

> So, sell Sugar to me!

I am sorry if I have in any way misrepresented myself.  I am not
interested in selling sugar.

On the other hand, I am interested in locating and fostering communities
which share common goals with Sugar and Sugar Labs.

Why Sugar?  I do not believe that Sugar is the one true desktop.  I
won't even go so far as to say that Sugar is the one true educational
environment.  What I do believe is that the open source development
model can be used to create an educational stack which is socially
beneficial and commercially viable.  Sugar can be an important part of
that stack.

Why Sugar Labs?  Currently, Sugar Lab's primary mission is to support
the One Laptop Per Child project.  Nearly all of our developers are
working to support OLPC either by directly supporting the current Sugar
release or planning future releases.

>From an economic perspective, the strength of the open source
development model comes from the ability for potential competitors to
collaborate on common frameworks while competing by differentiating
other parts of the stack.

OLPC is shouldering the majority of the cost of Sugar development.  My
goal for the last several weeks has been to help form communities that
share an interest in distributing the Sugar desktop to a wider audience.
The first step in that goal has been to encourage Linux distributions to
package the Sugar core and activities.

Once the packages have stabilized, I will recontact the educational
'spins' about making liveCDs and spins that directly support Sugar as a
desktop.  The third step in this process will be to directly approach
the distributions about integrating Sugar into their business models of
'give away the software and sell the support.'

The end goal of this effort is to make Sugar a commodity component of
the educational stack.  

Why embedded sugar?  The recent successes of the ASUS Eee PC, OLPC XO,
and Intel Classmate have blurred the definition of 'portable' computer
from shrunk down, ruggedized personal computer to embedded device with
extended IO capabilities.

>From an Engineering perspective, the goals of the primary target device
for Sugar align closely with those of embedded devices.  More with Less.
We constantly asd, 'How can we increase usability while reducing power
consumption, size, and cost?'

Traditionally, laptop manufactures have been competing by leveraging
Moore's law into more speed, memory, and features.  Consumers have grown
accustomed to the upgrade cycle.

One Laptop Per Child turned that model on its head.  They specified
minimum features that would meet their goals. They then designed a
device that would meet those goals while keeping cost and size at a
minimum.

It is reasonable to credit OLPC with establishing the $100 laptop
market.  Neither ASUS, Intel nor any of the other players would have
been willing to undercut their existing markets without the threat of
the XO.  While we are not there yet, the $100 dollar laptop is feasible
in the near future.

As with the Sugar desktop, I do not believe that the XO is the one true
laptop or learning device.  It is and can continue to be a valuable
hardware platform for running the educational stack.

Why embedded?  Current embedded CPUs are powerful enough to run the
Sugar desktop.

The embedded industry has years of experience designing and implementing
shock resistant, dust resistant, vibration resistant, and extreme
temperature resistant devices.

The embedded industry has years of experience competing on cost.

The embedded industry has recently made significant progress at reducing
power usage and extending battery life in cell phone and other mobile
devices.

The embedded industry is making progress developing toolkits where
porting software between platforms is becoming more and more
transparent.

I Haven't been working with Sugar on embedded devices long enough to
form concrete, long term goals.  But for the short term, and as a
possible presentation topic for the up coming conference, controlling a
LEGO mindstorms robot via Sugar activity running on a BeagleBoard would
be pretty cool:)

I hope this explains my goals for Sugar Labs and the reasons for those
goals.  If your goal coincide with any of our goals, I hope we can work
together for our mutually benefit.

thanks
dfarning