[Arm-netbook] iMX233-OLinuXino development started today

[lkcl luke]

On Thu, Mar 8, 2012 at 11:52 AM,
<arm-netbook-bounces at lists.phcomp.co.uk> wrote:
> The attached message has been automatically discarded.
>
> ---------- Forwarded message ----------
> From: simon <simon at mail.koala.ie>
> To: <arm-netbook at lists.phcomp.co.uk>
> Cc:
> Date: Thu, 08 Mar 2012 11:22:06 +0000
> Subject:
> On 08.03.2012 10:59, Gordan Bobic wrote:
>>
>> What's the point? For the same price you can get an R-Pi with am ARMv6
>> (vs ARMv5) and 4x the RAM.
>
>
> pi is an interesting project.
>
> they have taken a very specialised chip which just happens to have an ARM processor in it.
> an they have spun this into the next best hope for the education of the next generation of developers.
>
> they are very good at marketing. i wish we had some of that.
> it is a shame that they did not do it with a more open chip.

 yes.

> and why is "small" so important?

 reduced cost.  the larger the PCB and the more layers it has, the
more it costs to manufacture.

> the pc-card form factor that we are using is a much better idea.

 it's concidentally the exact same size as chosen by them.

> i would go further and have no connectors on the end of the card and route all signals through the 68-pin connector. it would be cheaper and all.

 yes... and that is fine: anyone may implement an EOMA-68-compliant
CPU card that has no connectors at the user-facing end.

 however if you want HDMI, that's 19 pins.  if you want Micro-SD,
that's another 10.  if you want Audio, and you want it as analog,
that's 4 lines, maybe 5.  if you want it as digital then it's 5 lines
(AC97).  if you want full I2S that's 8!

 the more "features" you add, the more lines are required.

there simply isn't enough room.  so we had to pick "Bus" interfaces
and "Lowest Common Denominator" Interfaces that have been around for
at least 10 years and so stand a good chance of being present across a
massive range of SoC CPUs.

l.