See the Maximite link below, its pretty straight forward.
You've touched upon some of the key issues. But let me walk through them with you and perhaps we'll get to a similar place ...
"All 3 solutions run Skype, let you create and manipulate documents and play games."
You've outlined some desirable features of this 'computer.' Skype, documents, and game playing. There is a trap here which I've discovered and that is with price. If a computer is 'expensive,' and any new laptop, netbook, or desktop computer will fall into that category, it has to have things which justify its existence/purchase. On the other hand, if the computer does not cost very much, people often dismiss it as a toy. (We have many exemplars from LeapFrog and others which fit this category).
If you had the equivalent of an Apple II (but upgraded to some recent microprocessors) you could have a computer which was programmable (and games of the type that make you think versus fast twitch) could be made for it, and on which you could edit documents but not change fonts or any of that stuff necessarily. And it cost less than $50 and the only 'breakable' parts were the keyboard and the TV you hooked it up to, it might be a different story.
"An Android tablet would seem to be the best solution when coupled with App Inventor. Sadly App Inventor doesn't run on Android itself so you have to have a non-Android computer to do your development work."
This is one of those things that is an issue for me as well. When I tried to use Arduinioes for teaching programming you run into a barrier of getting two computers rather than just one. And sadly the ecosystem of windows, Linux, and MacOS make compatibility a real drain.
"I could of course setup emulators so she gets a Sinclair Spectrum and Apple //e and can learn just how I did, but things have moved on."
I would hear more about 'moved on' ? Do you mean that you have moved on? Or that computer science has moved on? The beauty of these simple systems is that they are simple. They teach you to 'think' computer and that is tremendously powerful when you attack more modern systems. Sort of like a lawn mower motor on a tube frame with a centrifugal clutch isn't really a 'motorcycle' (its a minibike) but it is a great teaching tool on the concepts involved.
One of the areas that could use help on today is graphics. Back in the day there were some great pre-packaged graphics chips with full documentation that you could use to attach to your microprocessor. Sadly those days are gone. The good news however is that nearly every 'intro to FPGAs' class has a part where you build a video chip out of an FPGA. And they are now as cheap as those old customs chips. I've spec'd out a design which is a dual buffered HDMI output with a sprite based HW cursor. Its surprisingly cost effective (parts wise), and since it doesn't care about HDCP compliance its actually pretty simple.
With an open access graphics subsystem, a modest ARM core CPU, some memory, and USB plugs for "disks" and "keyboards" I think we could build a system that works well for learning about computers from 'Wow' to pretty sophisticated data structures. But that is getting ahead of myself.
> See the Maximite link below, its pretty straight forward.
Not everyone has VGA monitors hanging around (they don't) and only being able to do stuff in one place (where the monitor is) is a significant handicap over a tablet or netbook/mini that can be used anywhere.
While your theory about pricing is correct, in this case the parents really do want to spend about $500 and they really do need Skype, documents and game playing as the existing many year old family laptop is in its last legs. So yes they could buy two things - a cheapo for the kid and a second family laptop replacement, but that is not an optimal route to go. (The parents already have laptops, tablets and smartphones between them.)
> I would hear more about 'moved on'
When I was young having a tty program that asked your name and printed "hello $name" was cool. Nowadays a kid is not going to impress a friend with that. Remember that positive feedback is very reinforcing. On the other hand having an Android app that asks your name in a gui and gives a gui response is so much more familiar. App Inventor has media players, camera recording etc which is so much more modern and expressive. I'd much rather see her making an app where you record yourself saying your name, and then it says "hello $name" with your recording and text to speech.
> With an open access graphics subsystem, a modest ARM core CPU, some memory, and USB plugs for "disks" and "keyboards"
That just won't work. They skipped the whole generation of computers made up of discrete bits (processors, displays, disks, keyboards etc). Instead normal to them is integrated devices (processor, display, storage, wireless network, input mechanisms) and you pickup and use wherever takes your fancy.
Unless the kid becomes a programmer, the experience won't be relevant. Now, you could argue it improves thinking skills etc etc, but bare hardware is really only one layer of the abstraction chain- the layer you are probably most fond of. Every other layer- wiring, circuit design, device engineering, physics, and so on- are no less valid, but you probably can't do all of them. This is because we have to pick and choose which layers of abstraction to spend our time in- which gives the most bang for our buck; which is most relevant. For most people, "how to use a computer" in the office appliance sense is the most valuable.
You've touched upon some of the key issues. But let me walk through them with you and perhaps we'll get to a similar place ...
"All 3 solutions run Skype, let you create and manipulate documents and play games."
You've outlined some desirable features of this 'computer.' Skype, documents, and game playing. There is a trap here which I've discovered and that is with price. If a computer is 'expensive,' and any new laptop, netbook, or desktop computer will fall into that category, it has to have things which justify its existence/purchase. On the other hand, if the computer does not cost very much, people often dismiss it as a toy. (We have many exemplars from LeapFrog and others which fit this category).
If you had the equivalent of an Apple II (but upgraded to some recent microprocessors) you could have a computer which was programmable (and games of the type that make you think versus fast twitch) could be made for it, and on which you could edit documents but not change fonts or any of that stuff necessarily. And it cost less than $50 and the only 'breakable' parts were the keyboard and the TV you hooked it up to, it might be a different story.
"An Android tablet would seem to be the best solution when coupled with App Inventor. Sadly App Inventor doesn't run on Android itself so you have to have a non-Android computer to do your development work."
This is one of those things that is an issue for me as well. When I tried to use Arduinioes for teaching programming you run into a barrier of getting two computers rather than just one. And sadly the ecosystem of windows, Linux, and MacOS make compatibility a real drain.
"I could of course setup emulators so she gets a Sinclair Spectrum and Apple //e and can learn just how I did, but things have moved on."
I would hear more about 'moved on' ? Do you mean that you have moved on? Or that computer science has moved on? The beauty of these simple systems is that they are simple. They teach you to 'think' computer and that is tremendously powerful when you attack more modern systems. Sort of like a lawn mower motor on a tube frame with a centrifugal clutch isn't really a 'motorcycle' (its a minibike) but it is a great teaching tool on the concepts involved.
One of the areas that could use help on today is graphics. Back in the day there were some great pre-packaged graphics chips with full documentation that you could use to attach to your microprocessor. Sadly those days are gone. The good news however is that nearly every 'intro to FPGAs' class has a part where you build a video chip out of an FPGA. And they are now as cheap as those old customs chips. I've spec'd out a design which is a dual buffered HDMI output with a sprite based HW cursor. Its surprisingly cost effective (parts wise), and since it doesn't care about HDCP compliance its actually pretty simple.
With an open access graphics subsystem, a modest ARM core CPU, some memory, and USB plugs for "disks" and "keyboards" I think we could build a system that works well for learning about computers from 'Wow' to pretty sophisticated data structures. But that is getting ahead of myself.