Slashdotted

[Dr. H. Nikolaus Schaller]

Hi Lothar, Jörg, Sean, Steve, and all others who are interested in a  
future GTA03,

I have also been thinking for a while in exactly the same direction.  
Making the development of the GTA03 a community activity. Supported  
and sponsored by individuals and companies who are interested in the  
results.

On one hand HW development needs a clear and close communication  
between people. On the other hand it needs a large number of  
supporters who keep their fingers out until they are asked. But  
synchronizing activities is much more difficult than with SW (using  
SVN or GIT).

What we IMHO therefore need is:

* a core team that works (at least semi-)professionally on it. I think  
the community is large enough to provide enough members with all  
expertise that is required (from mechanics over battery to RF etc.).
* a clear milestone plan as in every successful hardware development  
project
* a specification freeze at some point in the milestone plan
* openness to ask for help into the community to judge between several  
similar technical solutions for the same requirement
* funding of the project organization (e.g. we can set up a community  
funds or society or association or however the legal form has to be  
choosen). Funding levels could start at 5€ per year for students and  
go upwards for individuals and companies. And special services (e.g.  
managing the production of 100 customized units) could even provide  
more funding for the organization.

A word to all those who think Hardware can not be developed by a  
community should take a look at:

	http://www.amsat.org/

Building, launching and operating not only 1 but approx. 50 satellites  
in the past 30 years is definitively more complex than building an  
open smartphone.

Nikolaus


Am 05.04.2009 um 19:39 schrieb Lothar Behrens:

>
> Am 05.04.2009 um 18:14 schrieb Joerg Reisenweber:
>
>> Am So  5. April 2009 schrieb Lothar Behrens:
>>> Hi,
>>>
>>> I am mostly reading and sometime writing here. If it was useful or
>>> useless - I don't know. But anyway.
>>> Isn't it possible to also develop hardware collaboratively?
>> [...]
>> Hi Lothar,
>> nope this won't fly. It's basically the sw pov approach to hw
>> development
>> Steve mentioned in one of his current posts.
>
> I may read his post...
>
>>
>> Developing hw is more than creating a good looking schematics in
>> Eagle, and
>> tasks like layout are partially done by autorouter and the other
>> half is a
>> *close* *interactive* process between the layout gal, the EE guys,
>> the RF
>> guys, the ME dept, sourcing dept etc etc.
>
> It is indeed difficult. But otoh are many hw projects (http://opencores.org/
>  or http://opencollector.org/ for sample),
> or at least hw related.
>
>>
>> IIRC OM had some really nasty experience when outsourcing some
>> layout task.
>> Just because the layout didn't understand exactly what EE had in
>> mind when
>> creating the schem, and EE didn't closely check the work of  
>> layouters.
>>
>
> I don't speak about outsourcing. I have made similar experiences with
> outsourcing:
>
> An EE project (motor control) should be outsourced, but the schematics
> were sent by faximile!
> The result was about writing an application to compare the netlists to
> compare the then distinct
> projects (different wire names and the like).
>
> So don't split any EE project or work with different versions without
> a CVS or SVN!
>
> But giving development boards or mobile phone development kits would
> be an option to
> broaden the idea behind open mobile phone. Say, a GSM kit could be
> used for the carPC hobby
> engineer. And there are really GSM modules sold by other companies. (http://www.gsm-modem.de/
> )
>
> Then you have the control about your pcb design, but propably broaden
> your product palette.
>
> Not all developers need a complete telephone. But you could indeed get
> more value if the
> 'components' of a mobile device also spread the globe - as a
> development kit or separately.
>
>> for your Q about project files instead of pdf: OM is making money by
>> selling
>> hw, so there's not much sense in publishing data that doesn't help
>> EE guys in
>> community to understand the hw but instead is only needed for
>> production
>> purposes. In the end you can't do anything on a single-device basis
>> with
>> layout or schem proj data you couldn't do without it. Or are you the
>> guy
>> who's etching 8-layer at home and soldering uBGA by hand? ;-) You
>> can't patch
>> a ready-done 8layer PCB, no matter what your document files are (sw
>> POV on
>> hw!). And no company is going to invest in producing some dozen
>> proto PCB
>> done by "anonymous" community guys, without checking each and every
>> trace and
>> footprint again what in the end for sure is more work than doing it
>> inhouse
>> from scratch.
>
> I don't mean that you grab the prototypes blindly for your use. But
> didn't you think, the comunity
> will also help in hardware aspects?
>
> Maybe the devkit could be coubled with a contest, who develops the
> best mod or addon. Or as an early preview
> for developers of software (the display discussion for sample:
> Touchscreen Capacitive (was Re: OT: iPhone howto)).
>
> Did someone yet really implemented drivers for a multitouch display?
>
> Wouldn't it good to get one preassembled from OM to develop for it?
>
> Therefore a kit would be good. Also selling kits for parts only.
>
> Gerald: It is worth to publish the private post :-)
>
>>
>> Other companies tend to keep schematics closed to protect their IP,
>> so we at
>> OM at least don't want to give asian cloners a kickstart without
>> adding *any*
>> benefit for our customers.
>
> Do you really think, they don't get any value just from the PDF  
> version?
> Time will tell us this.
>
> Other companies think about using the same idea behind open source for
> hardware. It's because
> of one big issue today: The technologie changes so fast, that
> individual development is too expensive.
>
> Opencores as mentioned above is impressing me. Another group is going
> similar steps in automotive. Even yet
> closed and membership is propably very expensive, But there is
> movement in how to develop technical products.
>
> Lothar
>
> -- | Rapid Prototyping | XSLT Codegeneration | http://www.lollisoft.de
> Lothar Behrens
> Heinrich-Scheufelen-Platz 2
> 73252 Lenningen
>
>
>
>
>
>
>
>
>
> _______________________________________________
> Openmoko community mailing list
> community at lists.openmoko.org
> http://lists.openmoko.org/mailman/listinfo/community