The University of São Paulo's intent to join Openmoko development

Jon 'maddog' Hall maddog at
Wed Jul 15 02:11:44 CEST 2009


>Could you clarify a bit what exactly is the goal of this effort?

Hmmm, I wrote a lot before I realized that you probably meant by "this
effort" the "effort" of having the university join the Openmoko project,
but originally I took it as the greater "effort" of bringing out an Open
Phone that people could use and extend and wrote a lot on that topic, so
please bear with me:

I can only give my impressions of the Openmoko project, where it is and
what I think the community desires from it.

I note that I have already answered one person who seemed to think that
putting the software stacks on a phone from Samsung was a better use of
time and energy than designing open hardware, so apparently there are
(as in any large community) many different ideas and goals.  Please bear
with me.  Some of these goals may be personal to me, and not shared by

>Is it just designing a line of "open hardware" mobile devices (that's
>what it sounds like), and selling them to manufacturers?

I have always hated the word "just", particularly when it is applied to
a person  "I am 'just' a documentation person" or "I am just an end
user" has always irritated me, but in this case I equally hate the word
applied to this effort.  And while I know that you did not mean to
offend, "Just designing a line of 'open hardware' mobile devices" grates
on my nerves a bit.  I do not want people to understate or underestimate
what has been done by the Openmoko project so far.

As to "selling them to manufacturers", the concept of having a design
freely licensable to different manufacturers so they can compete to
produce the lowest cost hardware is something that is good, IHMO.

Having many manufacturers producing "Openmoko phones" means that the
quantities go up and the number of platforms that can run any of the
software stacks increases....also good in my estimation.

Chip producers "respect" you and pay attention to you when your design
sells 50,000,000 phones a year instead of 12000.

The handset manufacturers I know are not interested in "one-offs".  The
investment they make in bringing out a line of hardware would not be
paid back until 100,000 or more units are sold.

>Or do they also want to help improve the software, like kernel, fso,
>x11 drivers, etc for the Neo?

I assume by "they" you mean the university?  I have invited them to join
the software effort, and I think they are interested, but they could
have done that at any time.  They could participate in the software
activities, just as the rest of the community has joined.

One slight issue, however, is the cost of the phones in Brazil.  Due to
Brazilian import duties, which can (with both federal and local duties)
range up to 94% of manufacturing cost and shipping) make the phone
*cost* a minimum of 550 USD FOB Taiwan.  This is without any of the
expenses or profit made by the importer.  By having the phones designed
in Brazil, with the import duty on the parts being 6% the finished
prototype phone prices might be reasonable for a Brazilian to purchase
even given the higher prototype costs.  Through a university resale
program we might get a considerable number of USP (and other Brazilian
university) students helping with the software.  Shipping phones out of
Brazil to other countries should not cost any more than shipping from
Taiwan (with, of course, the possible exception of shipping to China

I do think that if/when the university starts working with the community
full swing that there will be a lot more of the university students
getting involved, simply because of the university's involvement.  And
Professor Zuffo has indicated that they have various software skills
(codecs, security expertise) that they can add to the effort.

One of the issues here is that the project has not exactly been focused
on one stack of software....ergo the number of cycles that it is taking
to get any one stack ready has taken a long time.  IMHO this is both a
blessing and a curse.  A blessing because the hardware and kernel are
"tickled" in many ways, making the kernel more robust in the long run
and a person has choice in the software stack.  A curse because instead
of one intense effort we have several somewhat coordinated efforts.

>My feeling is that the GTA02 device itself is in pretty good shape
>compared to the software it runs. So how will the community benefit
>from a GTA03?

Yes, I agree the GTA02 is in "pretty good shape" compared to the
"software it runs".  However, it is in "good shape" for the middle of
2008.....not necessarily for the year 2009 or even the end of 2008.

Eventually the tide will turn and the software will be "in good shape"
while the hardware is perceived as being "long in the tooth".

If you pay attention to the hardware lists you will know that the Glamo
chip is not the best in the world, so removing it and doing a bit of
redesign will both save manufacturing costs and may actually give better
performance and/or battery life.

The GSM unit, I am told, is no longer being manufactured, so current
parts are being purchased from existing supplies, which become harder
and harder to source.  Like IPv4 addresses, sooner or later those
supplies will run a community we address it now or we may find
no new phones in the future.

Manufacturers do keep retiring parts.  This is why some notebooks, while
they have the same model numbers, typically have mid-production changes
in parts that drive OS designers crazy.  Usually the manufacturer works
with the OS provider to "smooth over" this issue, but it does happen.

Many of you may be satisfied with tri-band GSM, but it causes confusion
in ordering, creates extra costs for stocking and creates some "dead
spots" when you try to use an 900Mhz version in a place that has 850 Mhz
and vice versa.  Quad band fixes this.

And of course the world moves on with 3G deployment, with faster speeds
and lower costs of service.

So any hardware that stands still in effect moves backwards.

In addition to all this, some of you would like to have 3D capability
and a faster processor. While I do not consider the FreeRunner
objectionably slow, others have voiced that opinion on this list.

After having my picture taken with about 500,000 Brazilians (perhaps a
slight exaggeration), I can tell you that they would like a camera in
their phone.  And the camera would be useful in barcode reading
applications, in determining the legitimacy of money and other

I would like to have USB 2.0, longer battery life and a way to charge
the phone without having to insert a USB cord carefully into the socket.
Maybe have some contacts on the outside of the phone that would easily
mate with some contacts in a docking station or carrying case to charge

Finally, I read the lists where the hardware developers were looking for
things like an SMT line to do projects such as "GTA02-core", a cost
reduced version of the phone and continue work on the GTA03, the next
generation phone which addresses some of the aforementioned issues.

I found and proposed this partnership.


So coming back to *my* "Uber goals", I am hoping that this joining of
the University and the Openmoko project will allow the community to keep
generating Open Phones, each one better in quality, more feature-rich
and faster to market than the one before.

I would like to see the phone licensed to many different manufacturers,
but to interest them we need working software and to show them a path
for future models in quantities more than 12000.

I would like to visit a component manufacturer and tell them that the
Openmoko designed phones accounted for chip sales of greater than
100,000,000 chips a year....and wouldn't they like to make their chip
"open" so we could use it?

I would like the phones to attract a community of millions of users, so
we can have hundreds of thousands helping to maintain dozens of software
stacks for different people's needs....and to learn from one another.

I would like to see a whole "line" of open phones, so that the newest,
fastest and most expensive has a lower cost to me because I can sell my
older unit to someone who needs a good phone (and effective thin client
computer) but does not have the money for the latest and greatest
hardware. The software still works and can still can be updated on the
older phone because it is "open".

I want the Openmoko phones to eventually help bridge the digital divide.

I want to walk down the street and hear people say "He is a smart guy,
he is using an Openmoko instead of an iPhone".

And I want my mother (86 this year) to recognize the name "Openmoko",
without having to hear it from me.


Warmest regards,


More information about the community mailing list