developing C/C++ applications for OpenMoko -- where do we stand?

charlescooke.59846614 at bloglines.com charlescooke.59846614 at bloglines.com
Mon Nov 12 19:20:41 CET 2007


Let me preface my comments below by saying that I wholeheartedly hope the
OpenMoko project succeeds beyond it's wildest dreams.  It's an awesome project.



> -- writing C/++ OpenMoko apps is very difficult to almost impossible.

>   BIG PROBLEM...



I agree.  When I look at the wiki page on how to develop
a simple 'Hello World' application I get very concerned .. am I missing something?
 Why is this process so complex?  Is there an easier way?  If so, where is
this posted?



http://wiki.openmoko.org/wiki/Application_Development_Crash_Course



> As an interested onlooker I want OpenMoko to not

> only survive but
thrive, I am concerned that this project may

> not be able to reach critical
mass.



Ok.  Look at how far Linux on the desktop has come in reaching this
goal.  Despite all the efforts, it has constantly failed to penetrate the
mass (consumer) market.  It's not easy.  That being said, it's not necessary
to do this for OpenMoko to be considered a success - for instance, success
could be an OEM-style model where individual development shops customize the
device for various vertical applications/market segments and provide their
customers with full-stack solutions.  But even to reach this customer base,
it's vital that it *performs reliably* and development is fairly straightforward.



One major advantage that OpenMoko has over Linux-on-the-desktop is that
many of the problems that plague the latter are not there - differences in
hardware.  The hardware is essentially set in stone (I hope).  So why not
use this advantage and come up with a development process that is, for lack
of a better word, opinionated?  Make it very very easy to develop on the standard
hardware.  Assume that the hardware will be there, and operate on that assumption.
 Let's not duplicate the build processes on Linux that require accounting
for an infinite variety of processors, graphics devices, monitors, etc.




> If what Turkay says about the OpenMoko development platform

> is approximately
correct, this project may be doomed.

> How can we be this far down the road
without enough

> documentation that a larger group of developers can successfully

> write C/C++ apps for OpenMoko?



Not doomed :)  

 

> On the other hand,
one might also say how can a mobile phone

> project be this far along yet
have issues with

> fundamental requirements such as making and receiving
calls?



Yes, please!  Again, make the development platform standard with
the standard hardware and build a fully functioning core of a phone that a)
works, b) makes assumptions that the standard hardware is present and doesn't
need to account for corner-cases everywhere, and c) provides a standard kernel
that allows access to all standard hardware in a consistent fashion.



>
One possible conclusion: developing a full-featured open source

> mobile
phone  exceeds the scope of what a loosely

> organized open source community
can accomplish.



I don't think so, but I do think that it's necessary to
be more opinionated about the phone, it's capabilities, and the method of
development.  We don't need thousands of developers going off and developing
5 different windowing systems, 10 different APIs for making calls, and essentially
duplicating efforts while creating mass confusion to end-user developers/consumers.




More information about the community mailing list