I have to say, I am somewhat puzzled by some of the comments in this thread. I do wonder if the world looks that much different where-ever people live, compared to here in little Denmark.
*) Slight detour: Wired network versus wireless:
- Wireless is problematic in many institutions with a high concentration of clients. The interference gets atrocious due to packet collisions, once you hit a certain number of routers and clients within reach of all the rest. The slowdown you may experience due to this problem can be anything but trivial, and even dual band wireless routers is only a stopgap solution.
- Many buildings are being retrofit with a network connection at almost ever mains outlet, or they are built in at construction time. So I have been reliably informed by a friend of mine, who owns and runs a sizeable electricians installation company.
- Most every wireless router you will get from your ISP for home use, will have a built-in switch for wired Ethernet in addition to wireless. For instance the one I have has 4 Ethernet connections for the LAN. I suspect many consumers ask for this for their gaming rigs, so it is cheaper to just give this to everybody.
*) Cost of laptop/netbook vs. Raspi:
- Raspi is easily replaced if damaged for any reason. Surely I cannot be the only one, who have fried an expensive motherboard due to connecting my contraptions to the parallel port while experimenting?
- Raspi is cheap enough that you can give one to a young person in need. While almost all families have at least one computer these days, many families *will not* be able to afford the second one just for Johnny. Having one of the sons in the house wanting to use the family PC for hours on end is not going to work. Competing for the single PC is already a problem in many households.
On the other hand someone like myself would be able to find a second hand monitor, and flat out *give* a complete Raspi installation to somebody. Even more important would be the fact that the Raspi wouldn't be seen as a 'real' PC, so there would be less risk of competition for it from the recipient's siblings.
If people have a problem with this argument, then I suspect they may not have grown up 'on the wrong side of the tracks'. I did, back in the 8 bit days, and it was only due to sheer luck and a singular event that my parents were able to afford my first computer. In normal circumstances I would have had to wait until I got a real job to be able to afford one. Today a young teenager could finance a full Raspi installation just by running newspapers.
- The major problem with traditional PCs in schools is upkeep cost, in addition to the higher initial outlay. Think forced upgrade cycle from Microsoft et al., in addition to time spent on administration and security updates. The forced upgrade cycle has apparently not been of any obvious advantage to the schools nor the pupils for at least a decade. It only exist to act as a compulsory Microsoft tax on society. The Raspi has the opportunity to become a standard onto itself, ensuring long term compatibility at the application level. And even when you do want to upgrade a Pi installation, the cost of doing so is minimal.
*) Unexpected Raspi advantages aka. potential game changers:
- OPPC: One Pi Per Child. You can give each student/child their own Raspi, just keep monitors and the rest in the classroom/at home. Solves a whole bunch of problems with administration, access control, maintenance and whatnot.
- ODPA: One Distribution Per Application. Stop thinking of the Raspi as a traditional PC. Instead abuse the low cost of SD cards and think of the potential for instant OS/distribution swaps by replacing the SD card. You don't give people a copy of the installation files for your custom application. You give them a link to your customized Linux distribution, including configuration tweaks and relevant support libraries. No library dependency h*ll, no hardware conflicts due to standardized hardware, no need for convoluted changes to obscure text files, no problems at all.
For the end user: Download interesting 'application' image. Dump OS image to empty SD card. Swap card. Power up Pi. Play with app (which starts automagically on boot).
(The developer would of course also want to document how to install the application in an existing Raspi installation.)
- OPPA: One Pi Per Application. If you develop a particular, stand alone application, you no longer have to choose between and lock yourself into one of the major platforms. Just develop for the Pi and tell people to buy one if they wish to use your app. The cost is low enough that this is feasible, unlike telling people to just buy a Mac to run a program.
For instance you could (might be able to?) develop support for ARM's Serial Wire Debug (SWD) protocol via the GPIO interface, suddenly turning the Pi into a stand-alone development environment for a silly number of embedded MCUs. The cost of the Pi would be less than, say, the PicKIT 3 programmer, and with the latter one would still need to expose the PC to events in the mad scientist's laboratory.