FC project direction

[Spacefalcon the Outlaw]

Thanks to everyone who responded!  It was reassuring for me to hear
that my engineering common sense thinking that a device like GTA02
ought to be less efficient than a true dumbphone was indeed on the
right track.  It's just that previously I have heard so many praises
for FR running QtMoko (or even SHR) that I was starting to doubt if
there was any merit at all to my Free Dumb Phone idea.

I do notice one common thread though: there seems to be a not
insignificant number of people in our community (and in the larger Om
community from which we spawned) who use the GTA02 because it is
literally their first and only phone.  Furthermore, it seems that
these people have never experienced any non-smart phone before, hence
they have nothing to compare against.

To those of you who use or have used a GTA02 as your first and only
phone: let me guess, you were lured in by Openmoko's pompous advertising
of it as a totally free phone, and then very disappointed when Openmoko
failed to live up to the promise when it comes to the GSM part, right?

With leo2moko fw the GSM part is now free in both moral (what we have
is no worse than what "they" had) and practical (we can see and
understand everything it does) senses, and even the binary lib
components of this TCS211 fw are no major obstacle as the
corresponding components are in full source form in the other version
(TCS3.2), and our gcc-built fw is getting close to proving that these
full-src versions from TCS3.2 are a workable substitute for the
binary-only TCS211 ones.

But now that we have confirmation that the FR is flawed in not one but
two ways (not just the no-longer-available status, but apparently it
never was the epitome of perfection it is sometimes made out to be),
we have justification to continue with the plan of building a totally
free phone that will be plain, rather than smart.

Another factor to consider is that building a totally free and
practically usable dumbphone is very much within the capabilities of
our small team (practically me, DS and another friend of mine who is
not on this list who does PCB layout for OSHW projects): the hardware
of a Calypso dumbphone is simple enough that I should be able to
produce a decent design despite being nowhere near a professional hw
engineer, we have a good working reference source for all necessary
parts of the fw, and we understand everything well enough to where
we'll be fully able to support the product and ensure that it really
works no worse than a standard off-the-shelf dumbphone.

We don't have the same confidence base when it comes to free smartphone
designs like GTA0x.  From my observations on the Openmoko list, it
seems that both GTA02 and GTA04 users (and sw developers like Radek)
have oftentimes come to issues which they couldn't solve, and they
would just throw their hands in the air - "I/we don't know how to
solve this" - and go on using their devices crippled as they are, with
the user just having to live with and work around some non-understood
problem.  I personally don't see this situation being much better than
the closed proprietary stuff, from a purely practical standpoint: in
both cases there are some annoying issues which you can't fix and which
you have to just live with and work around the best you can.

Thus even though the number of people who would enjoy having a free
smartphone is probably much larger than the number of people who would
enjoy a free dumb phone, a free dumb phone of 100% stellar quality,
reliability and usability is actually within reach for us, whereas if
we went with the free smart phone route, it is highly doubtful that we
would be able to produce anything much better than the current GTA0x,
which I hear is still not quite up to par after all.  So it looks like
my Free Dumb Phone idea is still on the right track. :)

Now let me address a couple of specific points:

David wrote:

> Incidentally I was hoping that the Motorola and the Pirelli would have a
> freecalypso option someday, but since I do not have the skills to contribute
> I respect a decision to not go there.

It may still happen, it's just that there are a couple of factors
adding delay and uncertainty:

* Right now we need to get the basic core of our fw working no worse
  than TCS211, and it makes no sense to start looking into those
  "difficult" targets until we get this basic part done.

* Once we are done with all of the foundation stuff and are ready to
  approach the end goal of a usable product, we'll need to look at the
  relative costs and difficulties of reusing Mot/Pirelli hardware vs.
  building our own.  I mean, if building our own hw turns out to be
  fairly easy and cheap, why would any of us want to bother with Mot
  or Pirelli?  But if building our own hw is going to be a long and
  expensive venture, getting our fw running on C1xx or Pirelli in the
  interim would once again become an attractive option, and we'll
  revisit it then.

Nick wrote:

> I know Mychaela you have no interest in 3G / 4G hacking 
> (understandably),

An important clarification is in order: the issue is not that I have
no interest, the issue is that I/we don't have a leaked copy of any
vendor's 3G or 4G source code like we have for GSM.  If someone were
to publish a leaked fw source for some 3G-capable chipset, including
the actual source (as opposed to binary libs) for a working commercial
3G radio protocol stack, I would reciprocate by taking a very serious
look into possibly switching from our non-3G-capable Calypso to that
other chipset - but I have yet to see any such leaks.

> Also the fact that no free GSM stack exists is a really grave state 
> of affairs, and changing that would be wonderful :)

A known-working, commercially proven GSM stack source has been publicly
available since 2004 (well before Openmoko!) in the form of the famous
TSM30, but it was/is set up to run on unobtainium hardware.  Then we
got a much newer version with the Peek release some time between 2010
and 2012 - that's the GSM protocol stack source we are using in our FC
project as a replacement for the binary libs version used in leo2moko.
The bold contribution of our project is in the taking of these available
source leaks and transforming them into an actually usable product for
freedom-minded end users.

But it's just GSM; accomplishing the same for 3G or 4G will have to
wait until some (ex-) employee or contractor of one of the companies
in that business grows enough moral fortitude to leak some source.

VLR,
SF