Wonder if this little beauty would be capable of emulating some of the classic computer platforms as well - which of course gives scope for another market for it :)
(And also an excuse to play Elite as well :))
Despite all the excitement and expectation encompassing the RaspberryPi, the most remarkable thing about this low-power credit card-sized computer is its price tag: little more than £20 for a fully functional system capable of, among many things, 1080p video playback and hardware-accelerated graphics. The British-designed Pi …
...at least to my knowledge, the GPIOs aren't buffered on the RasPi, and there's no analog I/O.
And, if you blow up an I/O on the RasPi (which, my understanding is that's easier), you've gotta replace the whole board. Do it on an Arduino, and you need to replace a $4 chip.
But, a RasPi would be a good supplement to an Arduino for additional processing power...
You can get - for a bit more than this - a Arduino shield and largely code-compatible ARM Maple already. If you don't want shield and software compatible, there are tiny ARM-powered boards with loadsa-IO ports exposed. If you want Arduino plus 400x300 VGA, there's the lovely Gameduino.
The Raspberry is optimised for getting stuff on a screen vs hardware IO for the Arduino/Maple.
Me, I want both :)
Well if they get RISCOS running on this thing then YES. It'll run Elite and even better... it'll run D.Braben's other scifi game that amazing Zarch (of which a demo was supplied with every Archimedes under the name 'Lander'.) Zarch is better knonw on non-RIISCOS platforms as 'Virus'.
A beautifull game with a 3D vector-drawn isometric landscape. One of those unique games that unfortunatly never achieved the some success as Elite.
In fact I wonder why none of these old scool games never get ported to cell-phones. I mean even the ZX Spectrum had unique games (Marsport, Dun Darach, Tir Na Nog comes to mind) or the C64 (Head over heels, Zaxon) These could easily be ported to todays cellphones without (ab)using too much performance. But to stay in the RISCOS world. I'd love to see a proper port of StarFighter 3000 or even StuntRacer2000. They where both totally addictive (as well as Zarch by the way).
Perhaps Mr. Braben could supply 'Zarch/Virus' free with every machine sold (as an incentive to potential customers). IMHO I think Zarch is the more appealing game instead of Elite.
I've been following the project for some time and am eagerly awaiting the on sale date announcement.
I was at School in the 70s and 80s and strongly agree that the change from 'computer science' to 'ICT' was good for overall 'computer literacy' but bad for development of the industry as a whole. If this bit of kit can generate anywhere near the level of enthusiasm that we saw during the 80's microcomputer boom then it'll be a good thing!
Assuming the USB is USB2 (someone confirm, please) then you could plug in a USB Ethernet adapter for the second port. Of course that does add £15 - £25.
It's also a shame that there are few USB ports. Keyboard, mouse, disk and memory stick or DVD drive, I'd have liked 4. One could plug in a USB(2?) hub, but that's ugly compared to integrating the hub chip and , say, 4 connectors. How much would that cost - less, surely, than the £5 you can buy such a hub for. I'm sure it wouldn't complicate the board, just (maybe) make it a little larger.
Even so, if I can buy one as shown for under £30, I probably will. If it winds up at £100, I probably won't. I'll probably hack it into / onto a monitor, tapping the monitor's PSU which can almost certainly supply an extra watt.
The whole point of the board is to get the price down, even saving 25p is worth it - read the article! adding another USB or Ethernet is the *wrong* idea, after all, what would be next on the list USB3? surround sound? PIO disk interface? firewire? thunderbolt? nope, I'm sure that even one Ethernet was carefully considered,
I understood the point about getting the cost down. In my dreams it would have boasted SATA but I fuly understand why it doesn't. I'm not an electronics engineer so I may be complertely wrong about the following, but I'll fill in my thought process.
A USB hub chip is obviously a stand-alone chip plus support components connected to a computer system by four wires. Usually it's packaged in a little bit of plastic and the four wires are the PC-to-hub USB A-B cable. Therefore I would have thought it possible to design this system's circuit board so it can be populated either with a single USB connector, or with a USB hub chip and a quad USB connector. If I'm right the extra cost of the cheaper version would be the price of the extra PCB area needed for the (missing) quad connector abd USB hub chip. It's bad idea if (a) I'm wrong and it's not possible, or (b) if the extra cost of a small bit of unpopulated PCB on the cheaper version is enough to put off a significant percentage of customers. I doubt the latter.
Anyone with more detailed knowledge of board design or economics care to make a well-informed comment?
There are two chips on the Model B board - the SoC and a LAN chip. The LAN chip is multi purpose and takes the SoC single USB and converts it to ethernet, and also provides a two port hub.
This is a pretty cheap option - higher hub count and the required connectors increases the board price quite surprisingly - you need the bigger 4 port USB socket, the hub chip, and the board space on which to mount them - space is money. All together they put the price up by over 10%.
This is a bare board with exposed electronics components and adding a USB hub is "ugly"? I have a regular computer that encloses all the electronics and two additional 7 ports hubs on my desk behind my keyboard. Pretty standard to need a USB hub or two on any computer these days.
On the other hand I have a Dell keyboard that provides a USB port for the mouse so if you only needed a keyboard and mouse and had a keyboard that provided a USB connection, then you wouldn't need a hub at all.
I did visit the RasberryPi web-site. A box to put it in is planned, though it won't be available on day one. Then it'll be a neat little (tiny) box with a USB hub dangling from it. Ugly.
I'll also be paying more for the hub separately than it would cost integrated into the Pi. But that's all argued over above. If the added cost of unpopulated PC board real-estate really is enough to deter sales of the cheaper one-USB version, then ugly it'll have to be. (It's not much real-estate. USB hub chips are tiny and since the single/quad USB socket is either/or, they'd share most of the PCB location). Anyway - since it's less than the cost of a couple of pints, the money (unlike the aesthetics) isn't an issue for me.
A box?! Buy a tin of Altoids mints. Line it with something non-conductive, and you have a very strong case for your little computer. Hell, mount it upside-down with a thermal pad from the CPU to the tin, and you could use the tin itself as a heatsink. Might be good for getting a couple hundred more MHz out of the processor.
This is on my wishlist.
In the early days of USB it was claimed that USB didn't need to be daisychainable because it's so cheap that every peripheral would include a hub. But that never happened, did it.
In my opinion, the correct place for a USB hub is in the keyboard, which can then become that holiest of holy grails: the universal docking station.
Unfortunately, that was what should have happened years ago. It's unlikely to happen under USB 3 as who in their right mind is going to make a USB3 keyboard?
But still, we should have been at the stage by now where the majority of keyboards had a pass-through port for a mouse, and hopefully devices like this will finally start the creep towards that....
I think the Raspberry Pi is a great idea, but I think the FiGnition - http://sites.google.com/site/libby8dev/fignition - is an even better one.
And it really works - my 14-year-old thinks that the most exciting thing is to program the Game of Life, in Forth, on a little black and white screen, on a computer he had to solder and test himself.
Sort of reminds me of the Z80 and 6805 kits that were around in the late '70's and early '80's.
In other words - simple but fun.
I'm quite looking foreword to the RaspberryPi kit, as it seems like a good dev platform, plus it's basic specs blow the BasicStamp and Arduino (and their clones) out of the water. Plus the price is right - unlike chip mfgs dev kits that cost as much as a cheap laptop in some cases.
It's a lovely little machine, isn't it. And children seem to take to Forth very easily.
Forth is very very different to most programming languages, and I find that complete newbies learn it faster than experienced programmers. I think it's because they don't being any previous baggage and misconceptions with them.
Yeah, I was that age 'back when the Speccy ruled'...
Also, a kit that you can solder yourself without special tools or reflow ovens is nice. ;-)
(Just because I can solder 0805 SMDs by hand doesn't mean that I enjoy it. Soldering through-hole on a nicely-laid-out board, though, can be a joy)
As for hacking... Well... Just read the Google groups topics on the thing. Weird stuff is happening quite often. ;-)
but try getting information out of them if you want to hack your own drivers or %deity% forbid write your own code to run on their chips and aren't prepared to buy thousands of chips off them.
Broadcom is a very secretive, closed shop for hackers unless you're prepared to buy tens of thousands of chips or breach some very nasty NDAs.
I really hope Raspberry Pi is open and we can all get down to the metal on it but if Broadcom is involved, somehow I doubt it will be.
Having said all that, I will be seriously looking into buying a board when it's released.
"I really hope Raspberry Pi is open and we can all get down to the metal on it but if Broadcom is involved, somehow I doubt it will be."
If you look at open-source hardware projects, they usually sink because the specced components aren't available any more, but the Raspberry PI is as damn near generic as you can get. Ethernet controllers and USB controllers are pretty generic. ARM may be a propietary design, but it's licensable and so heavily commoditised. This means that they should be able to revise the hardware as the market changes and use components from other suppliers without invalidating the existing codebase and forking the system (and userbase).
The only truly closed component is the GPU. There really is no such thing as a generic GPU on the market, and GPU technology is changing rapidly. Mandating Open GL makes any GPU generic, and this means that they can change the GPU later when the current model is discontinued or they can negotiate a better price with another supplier.
On the other hand, if the GPU's native APIs were available, you can be damned sure that developers would use them, and software would become irrevocably tied to the current version of the hardware, which would destroy the long-term potential of the project.