Some commentary on the new Openmoko direction, and a review of FSO
antisol at internode.on.net
Thu Aug 7 19:18:35 CEST 2008
I just wanted to say thanks for this very informative response! :)
Wolfgang Spraul wrote:
> Dear Kevin,
> thanks for writing your review, for us at Openmoko nothing is better
> than a genuine reality check from outside.
> Executive summary first: I understand that we have not done a very
> good job at communicating our software strategy, and I accept
> responsibility for that.
> In detail, let me go through some of your comments:
> > The ASU was a proof-of-concept image that combined Qtopia,
> Enlightenment and GTK
> I would say it's more than proof-of-concept. When Qtopia became fully
> GPL in November last year, we looked at it technically. Trolltech
> published great binary images for the Neo 1973 but we couldn't
> immediately use them because Qtopia ran on the framebuffer not on X,
> and we did not want to give up GTK+. By February, we had come to the
> conclusion that porting Qtopia onto X, retaining GTK+ support, would
> be technically feasible for Openmoko. Our initial estimate was 2-4
> months. So we started. Qtopia is an addition to the Openmoko software
> > "ASU" and 2008.8 are the same thing for the most part.
> > The GTK based 2007.02 line was obsoleted (or some might argue
> deprecated ...
> I would say neither obsoleted nor deprecated. Bringing Qtopia (and
> especially Enlightenment) in _TEMPORARILY_ broke GTK+, that's one
> reason why we didn't release more ASU builds earlier.
> Actually the real breakage of the GTK+ telephony apps came from
> Enlightenment, which forced us to replace matchbox.
> We also wrote a new launcher, Illume.
> Then we went back to fix the GTK+ part, adding the theme back in,
> making it easy to switch from what are now the default Qtopia
> telephony apps to the GTK+ telephony apps. That part is not yet
> entirely finished. When 2008.8 is released, you still will not be able
> to remove Qtopia with a few clicks, and switch back to GTK+. Neither
> can you have both installed at the same time. We are very interested
> that both of that works, if nobody in the community picks this up then
> we will. We are already on it actually, like I said fixing the GTK+
> telephony apps that got broken by the introduction of Enlightenment
> and Qtopia is something we do for a while already.
> > The ASU and Framework were announced at roughly the same time.
> Please don't read anything into this. Openmoko is a real open source
> project. What you may feel are two 'announcements' 'at the same time'
> are in reality probably several emails from people in different parts
> of the world that are _NOT_ synchronized to the degree that you might
> see this from the typical company. In many companies, all you will
> hear is the marketing department. So they always speak with one voice.
> At Openmoko, even full-time, fully paid employees have no restrictions
> whatsoever to post private opinions, blogs, etc. In fact we encourage
> them! Please factor that in...
> > To the community is appeared as if Sean Moss-Pultz had pulled a
> decision out of his ass to abandon the software that people knew and
> go with Qtopia instead
> I agree some people felt like that, and it's not good. Again we have
> no unified marketing message, and we don't want to introduce any kind
> of 'gag order' for engineers (a lot of them would leave if we would
> try that anyway ;-))
> I will try to do better going forward communicating our technical
> > The biggest unanswered question was "I want to develop an app for
> Freerunner, what should I start with?"
> Yes, big problem. Right now a lot of the new things we are doing are
> implemented in Python using Python-ETK bindings.
> Check out the sources here: http://projects.openmoko.org/plugins/scmsvn/viewcvs.php/trunk/?root=exposure
> Now, this is not the end of the story. You can also develop using GTK
> +, or Qtopia. We even have quite good Java support (Jalimo) on the
> Neo. And so on.
> We are fully aware that we need to provide an entry point that is AS
> EASY AS POSSIBLE to use. We are working towards that, but I think you
> need to give us a few more months until we have good tutorials, sample
> sources, development/installation methods, etc. At least you will be
> able to see it coming together alive, I hope it's not too stressful :-)
> > Mickey Lauer posted again, making it pretty clear that my
> assumptions [that ASU and Framework were complimentary projects] were
> You must have misunderstood Mickey. ASU and Framework are absolutely
> Behind the scenes it is all melting together. Luckily Linux is a multi-
> process environment :-)
> ASU means Enlightenment and Qtopia were brought into the Openmoko
> software stack. It will be released as Om 2008.8.
> FSO/Framework means many more dbus services will be brought into the
> Openmoko software stack, including applications using d-bus. It will
> be released later, maybe Om 2009?
> OpenEmbedded is holding all this together.
> > If the two [ASU and Frameowork] were to merge, it would be by
> community support.
> Not correct. Full-time Openmoko engineers are already working on this
> > Frankly, after the disappointing abandonment of the Neo 1973 device
> That is really unfair. Let me explain the background:
> First, there are full-time, fully paid people at Openmoko supporting
> the Neo 1973 passionately. Just suggest something that would break
> 1973 compatibility to Werner and you will see what I mean. And this is
> fully supported and backed by Sean and our investors, even though, if
> you only think short-term, it doesn't make sense economically.
> Now - even better - I would _LOVE_ to produce new Neo 1973 devices
> right tomorrow. Imagine we could have the 1973 as a low-cost option
> next to the Freerunner. Maybe 199 USD. Without glamo, i.e. with a
> faster VRAM bandwidth. For a long time raster preferred developing on
> his 1973 instead of Freerunner prototypes!
> But you know what? I cannot produce them! The way the whole mobile
> industry works is that new mobile chips are announced, go into MP, are
> being produced, are being end-of-lifed. All of this very fast. It's a
> big machine, making more than 1 billion devices per year, and it
> doesn't stop for anybody. Everybody in this industry is under constant
> pressure to use the latest chips, all under NDA and even tighter
> Intellectual Property protections than before. The older chips simply
> cannot be bought anymore. We believe we can change this, we believe we
> can break the cycle and establish a long-lasting platform. But believe
> me we are up against a massive, scary machine. So if I hear a comment
> like "disappointing abandonment" that indeed is disappointing to me,
> because I know we have people who work late-nights every day and
> scramble to find those components so we could produce more 1973. But
> the machine has moved on. Eventually we will get smarter, use only
> chips that are not only open, but also produced for a longer time
> span. And hopefully with the help of everyone, we get bigger too, so
> we find more partners to break the system.
> > Sean is also the CEO of a company ... to ensure that Openmoko turns
> a profit.
> > The financiers, investors, venture capitalists and shareholders are
> the real chiefs. I suspect that those people started getting freaked
> out by the Freerunner delay.
> Let me praise our investors a bit. FIC is a wonderful parent company.
> They give Sean 100% freedom to execute _his_ vision. Sure, the
> Freerunner was delayed, but there is nobody outside that could
> possibly create more pressure on us than we put onto ourselves. People
> were waiting for the product to ship. We looked at _EVERY_ conceivable
> option to speed this up. As I said above the opening up of the Qtopia
> sources under GPL was one such option, and we looked at it. As you
> said, we were hoping it would give us a way to sell the phones to more
> 'normal' end users faster. To get some money into the bank. Not
> because our investors forced us into that, but because we wanted to
> stand on our own feet financially so we could further pursue what we
> wanted to do. Something like the framework, and other things we are
> dreaming about.
> > In some ways, it's a gigantic experiment.
> Indeed :-)
> Thank you so much for coming along on the ride, keep your feedback
> coming! (let me know if my mail was helpful, I answer more questions
> if you have more)
> Best Regards,
> On Aug 6, 2008, at 12:16 PM, Bobby Martin wrote:
>> Kevin Dean wrote a great review of the FSO Milestone 2 release, and
>> some commentary about the new OM direction that I think expresses the
>> opinion of a lot of us. It's not really surprising, but it is well
>> If it doesn't make you smile, you're doing something wrong.
>> Openmoko community mailing list
>> community at lists.openmoko.org
> Openmoko community mailing list
> community at lists.openmoko.org
More information about the community