[Tinkerphones] New tinkerphone gadgets in Goldelico shop?

[Paul Boddie]

On Tuesday 7. November 2017 14.56.41 H. Nikolaus Schaller wrote:
> 
> > Am 05.11.2017 um 00:15 schrieb Paul Boddie <paul at boddie.org.uk>:
> > 
> > I know you've been doing work for the Pyra project, and I guess that the
> > OMAP5 board is what you were calling "Letux Cortex 15"
> 
> yes. It is a tiny OMAP5 SoC board with 2 or 4GB RAM and ca. 32 GB eMMC and
> thrww board-to-board connectors. It has USB(3), SATA, I2C, HDMI, MIPI and
> audio.
> 
> There is a plan to offer it separately with enough documentation. But first
> we need to get the Pyra into production.

From what I've read on the Pyra forums, I'm sure there would be some interest.

(I also wonder what the interest would be for an open version of the GCW-Zero, 
given that Michael/EvilDragon has expressed dismay over the availability and 
quality of that product and yet has no difficulty selling them when they 
occasionally pass through his shop. But perhaps he should take it easy for a 
while before considering new projects!)

> > and that there was also a "Letux Cortex 8" board. So it would be
> > interesting to know what your plans would be for those.
> 
> Well, the Letux Cortex 8 is a tiny AM3358 based Linux board. I had planned
> to spin and offer a second version of it, but now the PocketBone came out
> and is essentially the same thing (except that I had planned for eMMC).
> And the PocketBone comes at an unbeatable low price so that I can't even
> get the components cheaper.
> 
> So there is no future for the Letux Cortex 8 board and I have added the
> PocketBone to the shop:
> http://shop.goldelico.com/wiki.php?page=Product&product=1057
> 
> The Letux-kernel for the Letux Cortex 8 works out of the box for the
> PocketBeagle.

That is too bad as far as you being able to offer a product of this kind, yet 
it then makes me wonder whether the PocketBeagle might be able to act as the 
basis of other products. Or maybe this would be too much like "bricolage" as 
Sven called it.

[...]

> Well, in my experience it is easier to modularize on schematics level and
> have individual boards than designing a perfectly stable and long-lasting
> interface. EOMA68 seems to be outdated to me before it is published.

Although I do support EOMA68, I can see that some of the choices are 
controversial.

It is interesting to see that Olimex have made a proprietary "SOM204" 
interface (3 * 68 = 204, coincidentally) with lots of signals to cover lots of 
applications, but this might be sacrificing the accessibility of the 
technology for flexibility. In other words, it may not be a problem for Olimex 
to put out new SOM boards, but perhaps the complexity might be too great for 
people wanting to make their own boards or the carrier boards that would 
accept the modules. Maybe that is the point of the exercise, however.

[Single-board computers]

> Well, they are more like high-level evaluation boards. Most of these boards
> are supported by the silicon vendors of the SoC. Their purpose is to make
> it known to a broader public (of engineers). So the SBC doesn't need to be
> perfect for everything.

Agreed. The big change was that such boards became cheap, perhaps by 
pretending that they could be used by a wider audience, instead of remaining 
niche items costing hundreds of dollars.

Paul