gta02-core (was Re: OM future)

[Carsten Haitzler (The Rasterman)]

On Thu, 25 Feb 2010 11:28:18 -0500 (EST) Ken Young <rtm at cfa.harvard.edu> said:

> Wolfgang Spraul <wolfgang at sharism.cc> wrote:
> 
> > If the FreeRunner would be bug free, I'm
> > sure people would still use them in 10+ years, easily.
> 
> The truth is that even though the Freerunner is buggy as hell,
> some people will still be using them in 10+ years.   Face it, we are
> now in the same boat as the Apple Newton fanatics.   We play with
> phones, because we enjoy playing with phones.   There is no viable
> business model here.   None.
> 
> Openmoko Inc had a far better shot at success than any open phone
> manufacturer will have again for the foreseeable future.   When the
> Neo 1973 came out, there were no mass-market linux phones available.
> The competition was relatively weak back then.   And still Openomoko
> was not able to make a go of it.   Perhaps better management would
> have made a difference.   But that doesn't matter now.   If you go
> to the maemo IRC channel these days, it's like an Openmoko reunion.

oh indeed. the community has dwindled to like 5% of what it was - as the 95%
(numbers out of arse again) are not as fanatical about having schemtics, cad
designs and every single thing open - they want mostly open - open enough so
closed doesnt get in the way of making apps and doing development - modifying
even the base system (to a reasonable extent). thats what the vast majority are
happy with. closed OpenGL-ES drivers become an issue to them - as they
sometimes crash or dont work and there is little hope of debugging your app as
its linking to a closed blob. they care about the practicalities of open
source. i'm in that boat myself. :)

> The people who want to make applications for a linux phone have
> moved on.   They were a significant part of the Openmoko community,
> and they want a linux phone that works.   If a gta03 were
> to go on sale tomorrow, I believe it would sell more poorly than
> the gta02 did.   And the gta02 didn't sell enough units to keep Openmoko
> Inc in the phone business.   The gta02-core and gta03 are of interest
> only to a proper subset of the people who were originally interested in
> Openmoko phones.
> 
> As Raster has pointed out, the idea that a group of hobbyists is going
> to make a viable phone in their spare time, using parts which are
> collecting dust in Sean's closet, is risible.   The other

well ok - you're harsher... i just wanted to inject reality into it - that it
wont happen without a lot of money - or with hyper-expensive hardware that is
probably well behind the existing "out there" mass market products

> possible future for OM software is anti-vendor ports.   It's hard to
> imagine that the OM software running on something like the Palm Pre
> will work more smoothly than the same software running on the OM
> hardware, for which the developers did not have to reverse engineer
> many things.   So if the anti-vendor ports are successful, we'll
> end up being able to turn something like the Palm Pre into a buggy
> hobbyist toy like the FreeRunner.   Hooray!

ahahhaha :) tho palm pre would be relatively easy - it's openembedded. its
omap3 where most of it is documented and open (excluding 3d unit). but... i get
your point :)

> Some in the OM community seem to suggest that if vendors *just knew*
> that they could have the wonderful SHR software for free, they'd
> design phones around it.   I disagree.   There is very little incentive
> for vendors and telecoms to support open systems, and plenty of reasons
> for them not to.   From a vendor's point of view, selling a phone
> that the user has full control of is a nightmare.   If something
> like the SHR stack ever actually entered widespread use, it would
> be the perfect platform for malware.   Users would be bricking their
> phones right and left.   Calls to service centers would go way up.
> Phone networks would be subject to DOS attacks.

well that's a bit extreme - as technically android and the nexus one or adp1
are pretty much this too. but you are right - this is a big fear inside
companies - you can rave on about open and benefits all you like - they are
scared of this kind of scenario and the bad publicity that comes along with it.
you need to loll them in gently one little thing at a time so they gain
confidence. let them get a little more open, see that it has no big bad side
effects - of anything positive ones, then the next small little thing. tryin
big-bang to do everything from hw to sw open is going to net you no love from
an existing maker. doing your own i dont see as viable - so you need to take a
new strategy. as i've explained already. :)

> The only reason that PCs are general purpose computers is a historical
> accident - PCs grew out of the hobbyist market, and hobbyists wanted
> a machine they could program.   There was no threat of external malware
> in the early 1980s.   The constant fight against PC malware is the price
> we now pay for that heritage.   Most PC users would be better off with
> a machine that came with a web browser, a few tools for photo manipulation
> and multimedia, and which could not have any additional software installed.
> With smart phones, the industry has a chance to replay history.   They
> can make the platform closed, and largely prevent the whole malware
> nightmare.   They can reduce the universe of software configurations they
> have to support.   It makes sense for them to do that.
> 
> Sad as it is for us, the most sensible approach for phone makers is
> probably Apple's.
> 
> I enjoy playing with my Freerunners, and my Neo 1973.   Others do too.
> But be honest with yourself - these phones are a dead end.   At this point
> we are like the nut-cases who want to run linux on their iPods.
> 
> Ken
> 
> 
> _______________________________________________
> Openmoko community mailing list
> community at lists.openmoko.org
> http://lists.openmoko.org/mailman/listinfo/community
> 


-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
The Rasterman (Carsten Haitzler)    raster at rasterman.com