[gta02-core] The University of São Paulo's intent to join Openmoko development

Jon 'maddog' Hall maddog at li.org
Tue Jul 14 20:30:50 CEST 2009


Dave,

> Hi Maddog,
> 
> [added cc to gta02-core list]
> 
> Thanks for looking into this - it certainly sounds like an amazing 
> opportunity, almost too good to be true - what's the catch!  :-)
> 
I hope that you will find there is no "catch".  If you do think there is
a "catch", please tell me.
> 
> Do you know how Dr Zuffo sees the universities involvement with, and 
> relationship to the rest of the community - i.e. do they seem
> themselves driving the projects, becoming 'sponsors', or as
> contributing members of the wider community?
> 
I will copy Dr. Zuffo on this, but I will also give here my perception
of where the University fits in.

Dr. Zuffo's LSI organization has been quite active in the past couple of
years in designing "set-top" boxes for over-the-air broadcast digital
TV, different types of solar-powered WiFi routers and other electronic
devices.  He showed me quite a few of these when I have visited him in
the past.

>From this he formulated the idea of having the LSI organization
formalized, and has been finishing up the creation of his
state-of-the-art facility, including an SMT line that mirrors the one
used by a large cellular telephone handset company who will go un-named
at this point.  Trust me, the cellular telephone company has nothing to
do with this project, and the only reason I mention them at all is to
allow you to know that the SMT line is eminently capable of doing this
type of work.

I might also add that Dr. Zuffo's set-top box used the ARM architecture,
so he and his students are very knowledgeable about it.  His group also
has knowledge about Intel's Atom processors and other chip sets.

Many months ago I met Dr. Zuffo, showed him the GTA02 and mentioned to
him the work the community has been doing.  I believe he bought two
phones from iSolve, Koolu's Brazilian distributor, to do additional
software work.

Recently, when Sean's company changed direction, I approached Dr. Zuffo
to see if the university would be willing to host the project.

I certainly see them joining as community members, and I think that Dr.
Zuffo might encourage some of his students and professors to join our
ranks even outside of this arrangement.

I certainly see the university as "sponsors" of the project, in the fact
that it does cost money to run such an SMT line, to do some of the legal
work, etc.  I would like to find a way to help compensate them for this
work, to make the project truly self-sustaining.  Dr. Zuffo and I have
discussed government grants and other funding ideas.  Please see below.

But in the end, I would like to see the community "drive" the project,
and from my conversations with him, I think that Dr. Zuffo is on this
page too.  However, I will warn you that he is serious about this, and
will expect capable leadership.  From my experience with FOSS groups, I
believe that that the community is capable of giving this leadership.
> 
>  From my point of view, it sounds like they've got a lot to offer the 
> projects (both in expertise and facilities), and I think our
> community 
> would be stronger with them as members.  I'm assuming from your
> messages 
> below that (initially at least) they are happy participating in the 
> kicad / CC-SA licensed community process gta02-core has adopted so
> far?
> 
> 
While I have not discussed these issues with him in depth, I think he
would embrace the use of kicad and other FOSS tools.  The University of
Sao Paulo has had a leadership position in Free Software use.

As to CC-SA, as I mentioned above, I would like to see the project
self-sustaining.  Under Openmoko there were a series of things that were
paid for by Sean's sponsoring company, and when they shifted direction
things started to get a bit dicey.

With regards to my other project with the professor, we have discussed a
model where the design will be fully open, designed openly, and
licensable by manufacturing companies.  However, as the project moves
from prototype and pilot to full production, the design has to move to
manufacturing companies who would pay a royalty of 1-2 dollars (probably
with a cap on total amount) for the design.  This would go back into the
design process at the University.  "Small" quantities (for hobbyist,
researcher and universities) would be licensed gratis.  By "small", I am
probably talking 1000-2000 units, with other licensing available for
special cases.

In order to fund the Openmoko project, I would like to suggest that
*all* the things that Openmoko made "open" *up to this time*:

o circuit design
o case design
o circuit board layout
o testing issues.
o plans for future, etc.

be completely open and published as before.

But (for example) the "gerbers" be licensed with a small royalty (1-2
dollars per phone, with a cap of 500,000 to 1,000,000 USD) only if the
party will make *over* 5,000-10,000 phones.

This way universities, small companies, etc. can get started, but if the
producers start to make and sell large numbers of phones, the university
will make some money in royalties to keep the project going.  We could
subsidize this with grant money available from various sources.

Companies would need to spend millions in designing a new phone, so they
would not be offended in paying for a design that would save them money,
yet small companies, researchers and hobbyists would be able to use the
output freely, even the "gerbers".

All the royalties would go to the non-profit institute at the
university, earmarked for the Openmoko project.  I could see the money
used to reduce the cost of prototypes, subsidizing travel to meetings,
acting as scholarships for student members of the community to live and
work at the university doing jobs for the production of phones (and
gaining valuable experience), and other areas of expense where the
community agrees the money should be directed.

This is only an idea, and I have not discussed it deeply even with Dr.
Zuffo.  I would like to hear your ideas and have discussion on this.

I feel it is EXTREMELY important to make sure the Openmoko project is
self-sustaining.
> 
> While we're a long way away from producing quantities of 10,000
> devices,
> 
The 10,000 devices was a number set by the university as an upper limit
for any project.

In production of systems there are typically "prototypes", then there
are "seed units", which may include field test units, layered
application development units, marketing units, etc.

My other project, for instance, may need 1500 to 2000 (or more) field
test units.  Dr. Zuffo's line can manufacture those to save me from
having to negotiate a separate contract with a manufacturing company.
Of course I would have to purchase the parts, etc. and I would encourage
him to charge a small royalty on the design to large manufacturers,
helping to make his line self-sustaining.

> in the coming weeks and months the gta02-core project is hoping to be
> in position to produce a handful of prototypes.  It sounds like
> LSI-USP might be able to help us with this current design process, and
> production of the prototypes?  This could be a sensible starting point
> for Dr Zuffo and LSI-USP to get involved with the community, building
> a relationship that can grow as we get to know them, and they get to
> know us!
> 
I agree completely.
> 
> 
> All the best,
> 
> Dave

Warmest regards,

maddog





More information about the gta02-core mailing list