Time for a tablet ?

Today we have different motherboards (Pi, Beagle, Panda)
Support for old computers (Acorn / Castle)
A laptop (Lapdock)
An UMPC (Pandora)

But not yet a tablet :)

There are 2 possibilities :
Use an existing design as the Archos 101 XS (basically a Pandaboard in a tablet). http://www.archos.com/gb/products/gen10/archos_101xs/specs.html?#a

Make kits around the next generation of LCD screens from Chalkboard : http://www.chalk-elec.com/?p=1859
with wood (http://makezine.com/2014/01/07/how-i-built-a-raspberry-pi-tablet/)
Acrylic (https://wiki.openjdk.java.net/display/OpenJFX/DukePad) – perhaps the best choice.
Or a case made with a 3D printer.

For size issue, the best is probably to use a Raspberry Pi A, without some of the ports, to make a very thin tablet. One USB Battery, and voila.

Is there someone with all the needed tools to make a small serie of this kind of tablet ?

Nota: we don’t have Arduino board too.
But this one is comping…
http://arduino.cc/en/Main/ArduinoBoardTre

It seems to me that the lack of support for tablets is partially down to a lack of useful documentation (ie the A10 chip), but primarily because you cannot ordinarily hook a mouse to a tablet. Our UI is still reasonably dependent upon the facilities of a three-button mouse and we have no real API for scroll mice, never mind screen touch and multiple-point actions such as two fingers down then pinch out. And that’s just for starters…

http://arduino.cc/en/Main/ArduinoBoardTre

“Arduino developers get up to 100 times more performance with the Sitara-processor-based TRE than they do on the Arduino Leonardo or Uno.”

Yeah.
They are comparing a 1GHz Cortex-A8 board to a 16MHz ATMega.
Duh.

(^_^)

On a hardware level, it looks fairly similar to the OMAP3 (Cortex-A8, and SGX530 video).

Our UI is still reasonably dependent upon the facilities of a three-button mouse and we have no real API for scroll mice, never mind screen touch and multiple-point actions such as two fingers down then pinch out. And that’s just for starters…

RPCEmu works reasonably nicely on a MacBook that has a multitouch trackpad. One finger = select, two fingers = adjust, three fingers = menu (via a Mac OS addon). The difference is that a tablet needs to decide which of these touches is the ‘click point’ while a laptop has a mouse pointer to do that. However picking the first point sensed might do it.

Would something like a Kindle Fire or Kindle Fire HD do the job? They have an OMAP 4 CPU inside. However one stumbling block is lack of wifi – don’t suppose you want an ethernet cable trailing from your tablet.

RPCEmu works reasonably nicely on a MacBook that has a multitouch trackpad. One finger = select, two fingers = adjust, three fingers = menu (via a Mac OS addon). The difference is that a tablet needs to decide which of these touches is the ‘click point’ while a laptop has a mouse pointer to do that. However picking the first point sensed might do it.

I’m not so sure if that would be the ideal solution – simply because I think people are likely to aim their index finger at the desired click point, and for the second finger will most likely be using their middle finger, which is longer.

I’d suggest the left-most¹ point in a multi-touch situation being treated as the click point makes more sense.

¹ Configurable for lefties so that it’s the right-most point – and, obviously, left-most/right-most would be determined by the orientation of the device.

I am struggling to think of an interface which would not be absolutely dreadful for trackpads or touch.

RISC OS simply isn’t designed for an elegant touch interface.

The nearest I have got for Touch is a three button touch bar at the top of the screen where you change from Select to Menu to Adjust. Menu click automatically switches to Select once the menu is open.

However RISC OS would need to be more resolution independent otherwise this would be hugely frustrating.

A multi-touch trackpad ie MacBook) could do up to 3 finger presses for menu and adjust as pointer info is already provided.

I wonder about the practicality of “tap and hold” to open a menu. It’s obviously not a perfect solution (and doesn’t solve the problem of the Adjust button) but could be better than nothing.

Personally, I’d much rather see a real laptop implementation. A tablet, while cool, I can’t imagine being very much fun to use. RISC OS is like Plan 9, it really, really, likes a big screen and a mouse.

The lapdock thing is an impressive contraption, but I won’t buy as it’s not really an every day, sling in the man-bag device. An ARM Chromebook which could run RISC OS would get bought right away, and I’d happily donate to a bounty or Kickstarter project.

It might be possible to use a microcontroller as the touch to mouse translation for RISC OS, so as far as the OS is concerned it is addressing a mouse, that can be done with a Pi talking to a ‘helper’ chip, so no reason why the idea couldn’t be extended to another board.

It might be possible to use a microcontroller as the touch to mouse translation for RISC OS,

I’m not sure how well this would work in practice. Doesn’t the mouse just supply a motion value (+/-X, +/-Y) and the OS adjust the pointer coordinates accordingly? If so, how would you ‘fake’ absolute screen locations without knowledge of what the ‘current’ position is?
[plus MouseStep, plus anything that alters the pointer position, plus the constrain behaviour of error messages, plus…]

Technically, a touch screen act as an usb mouse. In single touch mode only…

Did alan gibson write touch scrren driver for a POS system??

Technically, a touch screen act as an usb mouse. In single touch mode only…

Do you have a reference? I have found protocols for serial touch screens, and incomplete details of the HID for USB ones. However all return the point press as a coordinate.
This isn’t to say that they are not “treated like mice” by the OS – it would be pretty simple for a driver under RISC OS to alter the pointer position and emulate button presses. But this is absolutely complicit upon the driver with the OS as it needs to be able to exactly place the pointer.
This is why I do not believe a microcontroller, external to the OS, could ever achieve the same thing.

Does RISC OS have any support for graphics tablets? I’d imagine it is much the same concept.

For example with this : https://www.adafruit.com/blog/2013/11/08/new-product-resistive-touch-screen-to-usb-mouse-controller-ar1100/

For example with this : https://www.adafruit.com/blog/2013/11/08/new-product-resistive-touch-screen-to-usb-mouse-controller-ar1100/

Follow a link or two and you end up looking at this and that page has this

i.e. an open source driver for linux. The little converter board is cheap enough, I wonder if Rick can blag a discarded touch screen to play with :)

I’m wondering about pulling the guts out of a Fujitsu Lifebook to see if a Pi will fit¹ – the touch interface might be handy.

¹ Could be ages, still working single handed at work.