Re: [Ghm-discuss] [gnu-prog-discuss] Planning the next GNU Hackers Meeting

[lkcl luke]

On Mon, Dec 19, 2011 at 10:39 PM, Christopher Allan Webber
<address@hidden> wrote:
> I would also like to see GNU move into this space.  We've seen GNU/Linux
> head down this path several times but flop and fail (and never
> completely free of course, often times not even close, but the
> infrastructure pointing in a direction that *could* have become free).

 one step at a time, yes.

 the problem is that because the economics are against software
freedom at the hardware level, any project which _could_ get FSF
backing has no FSF-endorseable hardware to work with!

>  - Instead of reinventing the distro, try to work *with* distros like
>   Debian, Trisquel, Fedora to get the mobile tooling packaged.

 my personal preference - purely for ultra-simplicity and size - is
openembedded angstrom linux with either gpe or the other one, it's
qt3-based.  you _can_ get phone applications *already* under these
linux distros.

 what i particularly like about these distros (the ones that run under
angstrom) is that they are absolutely tiny.  they date back 10 years,
and are capable of running under 90mhz CPUs with only 32mb of RAM and
16mb of NAND Flash.

 thus, if you _just_ want phone calls, they're ideal.


>  - Also, on the "tablet" front, I've been using Gnome 3 on my combo
>   laptop/tablet thinkpad x220.  It's *very nice* in tablet mode, and
>   thanks to the intel drivers it's possible to completely run a free
>   software system with this.

 yes.... and how much did it cost, as, say, compared to a $100
GPL-violating Android mobile phone with a Qualcomm MSM chipset or
there's another one which is also GPL violating and runs non-free GSM
Radio ROM firmware in shared memory between the CPU and an on-board
embedded Baseband Processor?

> We've seen a lot of attempts to move into the tablet/mobile space that
> have failed because it's hard, because it wasn't clear what the future
> would look like here, and because manufacturers have felt the need to
> reinvent everything.  Something along the lines of the design of meeting
> Gnome 3 and Maemo 5 halfway while using a generic API under the hood
> like freesmartphone.org provided would really fit that pretty well.

 really - it's too much.  the lesson learned from openmoko was *do
not* reinvent *anything*.  the openmoko project failed because they
tried to develop an OS at the same time as developing the hardware.
they failed because they didn't think "why don't i just grab this
pre-existing gnu/linux distro which has a built-in basic phone
application and WHOOPSIE we just learned, only 4 months into the
project, that there's howling R.F. feedback between the GSM Radio ROM
and the microphone".

 instead, it was i think almost 3 years into the project when they
discovered this problem, after spending 85% of the available budget,
and the critical chosen components had reached end-of-life!  they then
didn't have enough money - out of i think it was a $10m budget - to
redesign the hardware, finish the software and keep people interested.

the lesson is: _don't_ reinvent the wheel - ship whatever works.
people will be grateful that it makes phone calls, instead of
overwhelming them with features that they don't understand and don't
want.

 ok that's oversimplistic but you get the point i'm sure.

 l.

 p.s. the GTA04 by dr schaller of goldelico is, now, i believe,
shipping.  you can place an order for a GTA04 motherboard that will
fit in an openmoko phone.