Open Source Hardware
- Ed Prentice,
Thank you for your letter and your hardware experience which I share.
It's extremely helpful to get advice on hardware issues!
I think that you, Ricardo and I are not in disagreement. And that
I aspire to be flexible and not dogmatic. For me, "open source hardware"
is the ability, in theory and practice, for anybody to adapt part of the
design, as much as possible, to serve their purposes and situation.
I take this approach because:
* I don't want to invest myself or others in projects that will never happen.
* I'm happy if manufacturers can make Includers available and profit from
* I believe that the human assets we are assembling and unleashing are
much greater than that of any devices.
* I think that we are and will make progress step-by-step by tackling real
world challenges that we are facing in Africa. Ricardo and our African
team are a great inspiration.
* I think that our role is to break down the obstacles in to smaller
tasks, to find partners who might tackle some of them.
* I think that we should invest ourselves in looking for partners who like
our values and care about them and our relationship.
* I think that the skills developed in experimenting are very valuable for
our team, especially in Africa.
* We are stewards of an open source software and hardware system that
allows for a wide variety of Includers and solutions.
* I think that we will earn our money because of the value of the team we
are assembling online and on-the-ground.
Tactically, encouraged by Barry Dobyns, I have focused on the engaging
companies for whom we could make their online service effective in the
offline world. That's how we might get funding to design the offline
interface. Once we have an offline interface (say its coded in Python on
Linux), then we can run it on various devices, and then we have a much
better chance of getting manufacturers interested. In particular, I think
that there are digital picture frames which may perhaps already allow
engineers with the right back doors to do data entry using computer
keyboards through the USB port. In this sense, we agree, yes?
I don't have the business contacts to whom I might suggest the market
sizes and optimal prices and I don't have the data for that, either.
However, I do think we should interview our participants so that we know
exactly how they are accessing the Internet, how much that costs, how can
we improve that, because then we are helping them directly and also we are
getting the data for the guesses about the market that we can make in the
Ed, Thank you! and I look forward to more letters from you and others, too.
Do you have any contacts in Nokia or other companies that I might
approach, especially while I am still in the San Francisco area, until
Andrius Kulikauskas, Minciu Sodas, http://www.ms.lt, ms@...
I did spend some time thinking about the Includer design objectives. Upon
further review of posted documents, I don’t think it would help right now
to help define the problem further. I think you have moved way past that.
(If you have not already, please consider all the variations of the USB
stick. MP3 players and camera memories serve the same function.)
Here is what I believe is your biggest challenge and what I am concerned
about. Understanding hardware production is non-trivial. Most outside
observers frankly have no clue how hardware comes to be. I have a long
experience in hardware—and that is why I am working in software!
As I mentioned to you I have concerns about trying to separate social
non-profit initiatives from commercial initiatives. I try to see the good
value that that emerges from whatever the source. I think embracing all
possible “components” can get effective results easier and faster,
regardless of the motivations of each component. The relevance and value
is in the solution obtained not so much how every dimension came into
Specifically I keep stumbling where it seems that the Includer concept is
politically and philosophically encumbered by social engineering concepts,
while I realize the objectives are to free it from such constraints. I
realize you are guided by very strong personal philosophy that I haven’t
completely understood (Many find me complex but I seem very primitive in
comparison.) The open sourcing concept of hardware is far more than a bit
challenging. Hardware cannot just appear but is a factory driven activity
that to be effective needs a large economic base/market. Your goal should
be to discover how to exploit this driving engine. There is no other way
to be cost effective in world markets today.
The best open source hardware example that I know of is the original IBM
PC architecture which continues in variations; and a particular the PC104
architecture which is an industrial computer hardware concept built on the
IBM design. The PC 104 concept (and other industrial computer
architectures) offers a flexible, expandable hardware architecture.
Theoretically it is ideal for projects such as Includer. Unfortunately,
this cannot be expected to be a low-cost solution in the context of
serving developing markets however. (There is also a new platform of
handheld devices that has a similar component architecture but I don’t
believe it is applicable to a low cost market. Sorry I don’t recall the
The OLPC project should have demonstrated these challenges for anyone.
There is simply too much friction along the path to the end objective.
Frankly, the social objectives cannot overcome this friction. Software
development and intellectual processes fit well, but factory production
required for low-cost hardware must have an economic basis of large
My suggestion is to closely monitor and analyze existing platforms. Find
products that are driven by volume manufacturing needs. Identify what
elements are missing if any. Analyze the weaknesses and determine what can
be done with the existing functions. When possible candidate platforms are
found consider what hacks may allow them to better be deployed. It did not
take long for the iPhone to be hacked to open up aspects of the design,
I believe you will be served far better by looking for high volume
consumer products that can be adapted—in some dimension. The more
physical/hardware modifications required the less likely a solution is to
be workable. A hobbyist mentality is to some extent the enemy of finding
the best solutions. I would think that a big part of an effective solution
may be finding sources of excess inventory and retired products. The world
is swimming in old mobile phones. Consider this product lifetime flow.
There may be a very green recycling solution for some class of device
If I had this problem to solve I would buy a Nokia WIFI device. The Nokia
770 sold for under $150 as it was phased out. It is an amazing device that
as far as I can tell would solve the core problem you are identifying.
(This is an ongoing product line, now 800, 810) In this regard I would try
to include wireless capability in your design alternatives, Bluetooth and
WIFI links (maybe others) can eliminate hardware hacks that pose some risk
to reliable solutions.
I would recommend making a connection with Nokia while you are here. You
may also consider a Palm connection for both the hardware and software
Leaping to what I would do:
Define some market requirement in terms of cost/price and size of market.
This will enable the evaluation of potential platforms, and maybe the
timeframe for when they can meet an objective. I think time is a huge
variable for this project. (reference the OLPC evolution) How much would
it cost this year, how much can some one pay this year, next year, etc. I
think you have already stated that for this initiative to work it should
not be dependent on a government type of subsidy.
For candidate platforms you would want a technical evaluation (This may be
the single most important thing you can do):
Processor MIPS etc.
Memory – internal/pluggable
Retirement expectation and value
Ongoing family or single product
I would hope for multiple intersections of various platforms becoming
available and a market able to acquire at various price points. Depending
on the timing and degree of difficulty you could then work to match
inventory opportunities with need.
I did not want to post this to your wiki/list. It may seem a bit divergent
which I do not want to be. There are process issues in the creative
process that cannot be opened to random input—this is why I suggested that
someone should be in charge of a task and manage it. (Cat herding is a
dangerous game in actually creating something, which is very different
than ideation which is a specific phase of design.) I do have enough
expertise in the volume hardware markets that I did want to contribute my
view on the implications. I find that approaching hardware requirements
without appreciating the challenge is extremely difficult—i.e., unlikely
to yield an effective result.
Hope this has been helpful in some dimension.