Slashdot Mirror

Link

https://beagleboard.org/e14-bb...

As an example, and with its own problems, the Raspberry Pi has proved "good enough" for a lot of embedded companies, who just accept various tradeoffs to code in Python (or whatever) under Linux. In ten years, such hardware will be even cheaper and even better, and may have even better RTOS support:
http://pebblebay.com/raspberry...

Likewise and even more so for the truly free and open source Beagleboard family:
http://beagleboard.org/

What a far cry from the Kim-1 with 1K of memory I bought as a kid (with my father's help) for probably 50X in current dollars what a Rapsberry Pi or Beagleboard costs and supplies about 1GB memory. We even had to build the power supply ourselves. :-)

This is not to disagree with what you are saying right now. I know how hard and important all that work can be. But if better tools eventually let fewer engineers produce more projects in the same time, and cheaper hardware means less constraints, and other standards change expectations so customers know to work within them, than the need for managing such complexity (including the human side) may go down. Granted with a rising increase in an internet of things and robotics, embedded work may well still increase in demand for a decade or two until better standards come to dominate the field and change the nature of so many embedded tasks. So, for anyone already well enmeshed int he embedded field, it may well be a good gig for the next decade or two.

Anyway, for fun, to go through your list (a bit tongue in cheek, so not completely serous answers):

"First of all the robot would have to sit in meetings with the customer understanding what the customer wants, telling him what can't be done and outlining alternative solutions in dumbed down language the customer understands. It should tell the customer if some of his choices will raise the cost (Yes, in theory the hardware can do X but the current drivers can't)."

This can be replaced in part by a web interface where the customer clicks on options and the software tells them if the combination is allowed.

"It would have to write an offer listing everything that the customer wants to have implemented but worded so he can't expect more for his money. It has to be worded so that sales people and management understand enough to agree to the price written at the bottom."

Again, web backend generates this based on web interface choices.

"It needs to understand all documentation provided by the customer and it needs to be able to find more. It also needs to know who it can ask for an undocumented detail. Currently documentation includes data sheets, Doxygen like API descriptions, articles, standards, schematics, forum discussions, RCS commit messages, source code, and books but there will probably be a new form designed to be understood by machines. It is also useful to remember everything the customer said even if it might turn out to be wrong."

This could be mostly replaced by machine-readable standards as you suggest.

"It has to know sources of errors. Documentation can have errors. There can be errors in the design of the hardware. The hardware might be faulty (f.ex. bad solder joints) and you need to know what will destroy the hardware (no you can't configure that GPIO to output a low value). It has to fix simple errors itself (yes, I did have to solder some jumper wires and pull up resistors in the past few years). If it can't fix the error, it has to discuss the problem with someone who can. If the problem won't be fixed (no we won't spin a new SoC revision for you) software workarounds have to be devised and the pros and cons have to be discussed with the customer. To be able to find errors, it has to know about software and hardware debugging techniques (printf debugging, Gdb, Valgrind, JTAG, oscilloscope probing, ...)."

Bad hardware becomes so cheap it is discarded. Twenty years ago O

Good catch! OLPC lost a lot of developer mindshare IMHO when they started cosying up to Microsoft and changing their hardware to run Windows. Example:
http://www.olpcnews.com/softwa...
"For me, that paragraph represents the end of a dream. I say that XP on the XO is the end of One Laptop Per Child as an educational project. With a Microsoft operating system, an XO becomes a "$200 laptop", a cheap Toshiba replacement, not an educational learning tool for children. With the Sugar User Interface, OLPC can claim to have a Constructionist learning methodology, it can claim to be promoting exploration and learning, it can even hope to activate the view source key. But once you put on XP, no matter how much it may be customized to leverage the XO hardware, children will not be taught to "learn learning" as Negroponte promised. They will be taught "ICT skills", a phrase Negroponte himself railed against. Ministries of Education will be tempted to lock down XO's in computer labs and revert the whole one laptop per child idea back to one to many, effectively negating the goal of this grand dream. Yes, for me XP on the XO is the end of OLPC, no matter who is the CEO."

Hope Raspberry Pi does not suffer the same fate -- especially as I recently bought two B+ versions, :-) not knowing about either of these forthcoming changes (better hardware or Windows).

The last week or so, I've been watching for the new Beagleboard-X15, which is both open source hardware (Raspberry Pi design is not quite open hardware it seems) and will answer a lot of performance and memory issues at least compared to the Raspberry Pi B+ or the Beaglebone Black.
http://elinux.org/Beagleboard:...
http://beagleboard.org/project...
"The BeagleBoard-X15 is the newest member of the BeagleBoard family. Measuring 4" x 4.2", it is based on a Dual Core A15 processor running at 1.5GHZ and features 2GB of DDR3L Memory. It is in the beta phase. ... Guidance is that it is certainly over $100 ..."

So, that board is a lot pricier than this newer (or older) Raspberry Pi though. Not too much for a typical home office server use as an example (like to run NodeJS locally for testing on a separate non-VM box), but still 3X to 4X more for the board. However, when you add a case, extra media like a hard disk or big USB flash drive, and a power supply, and a wireless dongle, and so on, I doubt the overall cost is probably that much more than 2X for an entire system with the Beagleboard-X15.

Most of the competitors to the pi are junk. They're unsupported, not open source.

The Raspberry Pi is not open source hardware, it never has been. You can't even get the gerbers nor full documentation on the Broadcom BCM2835 SoC, let alone the full design files.

In contrast, competing boards like Beaglebone Black and Olimex's OLinuXino range like the A10-OLinuXino-LIME, A20-OLinuXino-LIME and A20-OLinuXino-LIME2 are fully open source hardware with all the information being provided. In the best tradition of open source, everyone is welcome to make their own derivatives using these open materials. All four boards are also substantially more powerful and flexible than the RasPi, very well made and fully supported.

The Raspberry Pi Foundation has declined to make the RasPi open source hardware despite years of requests from the community.

I had a beagleboard http://beagleboard.org/beagleb... , $125, though about £120 GBP to order in the UK. So switching to the pi gave me something smaller, cheaper, more reliable (I had various issues with the USB on the BB), with ethernet and accessible GPIOs.

Beaglebone Black is quite nice, and closer in price to Pi, but I have not personally made any projects where it would be an advantage over a Pi. I for the Pi that I use as a desktop (mostly with just a bunch of terminals SSHed to big machines), it might be a bit faster, but I'd need to add a USB hub to connect a mouse and keyboard.

I have a couple RPis and they are fun for out-of-the-box projects like RetroPi, but it's a BBB that I trust to run my 3.25hp router around my CNC table at 200in/min. Though, recently with the work going into MachineKit that's pretty much an out-of-the-box project too.

RPi had quite a bit of energy in the community to begin with and that momentum still persists and give a bit of an advantage to them in project development, but that will only go on so long with the arrogance displayed in that thread. Don't venture into the OSS space and start complaining about derivatives of your work. Especially when the project is mainly sponsored and developed by employees of the company that is selling the bloody chips that are nowhere near open or documented while all the function of your system is dependent on an OS you did not develop or pay for.

http://beagleboard.org/black

It used to be $45, but it has apparently gone up to $55 :/
Although, it still eats RPi lunch for anything other than being able to play media. (RPi VideoCore can flick 1080p hardware-decoded video, the BBB can't.)
Tons more GPIOs, over twice the real-world performance in computing, and 2 dedicated processors to run deterministically timed code for GPIO/peripherals (outside of the OS)

How does the ballpark cost of $600 arise when this device is essentially a BeagleBone Black in a box with added LCD and keyboard?

What's more, BBB is fully Open Source HardWare too, so even the user freedoms that OpenPandora wishes to provide are maintained.

What does the extra $550 give you?

I was thinking of things like the olinuxino which is cheaper than a Pi and seemed to be building a fanbase when a change from mailing list to web forum chilled the community. Or the Beagle Bone which is more open and technically superior, but has a nasty habit of throwing up problems no one else seems to have ever seen (according to google, that is). Or the Fox G20 which is a nice board and seems to have a good community, but isn't generally well known since the bulk of the community round it is still Italian.

ive never cared for the Pi for a few reasons, call me a hater but ive reasons..
1. power: this thing is emaciated by any standard. its got plenty of connectors but driving a media center like XBMC is a chore. with 15 watts of power i can trounce this thing with an Atom.
2. its encumbered, so enjoy one more $45 product that kicks FLOSS to the curb.. Beaglebone isnt encumbered...but beagleboard isnt the word for god on the lips and hearts of every blogosphere hipster.
3. it has no practical use. wireless access point? for 5 watts more you get better antenna gain, a better transciever, and PoE in a $35 tplink package. media center? enjoy the one skin for XBMC you can actually run, and the analog audio for icecast streaming will grind it to a halt.

i like low-power embedded stuff, you should too, but there comes a point at which it needs to do something otherwise its just more electric garbage. if you want to run a kegerator with it, or a home automation server, then i still suggest beaglebone http://beagleboard.org/Products/BeagleBone

TLDR: hipster blogcred articles make my shit itch.

You listed an Android tablet and two Android sticks. I think you missed the part where he said "as good as Pi", since none of those fit that criteria.

What you should have said was BeagleBone Black, $45:
http://beagleboard.org/Products/BeagleBone%20Black

When did TI buy the Beagleboard and start selling it?

http://beagleboard.org/ is the REal thing and they dont seem to act like TI bought them.

And I cant buy one from TI 's store, it redirects me to Beagleboard.com

The Raspberry Pi is not open hardware at the board level (schematic but no gerbers) nor at the SoC level (no full reference manual on the Broadcom BCM2835 device) nor at the boot level (booting and boot options are handled by the proprietary VideoCore IV) nor at the GPU and DSP levels (the VideoCore is entirely closed/under NDA). In fact, the only fully open thing about Raspberry Pi is its old and rather obsolete ARM11 processor.

So why exactly is anyone associating the word "open" with Raspberry Pi?

Far more open is the similarly priced BeagleBone Black, which provides full gerbers, full SoC reference manual, and full open source boot control (U-Boot). The BeagleBone Black's TI SoC does have a closed GPU, but since the board isn't aimed at running games nor consuming media like the Raspberry Pi is, it hardly matters. And the BeagleBone Black is far more capable in almost every other respect.

It's cool that Raspberry Pi has helped to bring ARM board prices down, but it shouldn't be called an open platform when it's mostly closed.

The same BeagleBoard-xM, yes. The BeagleBone Black has different specs.

I am not saying the product is without merit. I do like the form factor, but right now I have two Raspberry Pis, a BeagleBone, and a pcDuino on my desk (for use in various client projects). Those are just three of the various hobbyist and industrial small ARM based systems out there.

Right now the EOMA-68 is more or less vaporware. Wake me when I can buy one, then we can talk...

The RPi might be a bit underpowered for what you want to do with it, especially the Android idea. And IIRC, they even haven't been able to port Android yet.

Better choice (albiet more expensive) would be a Beagle Board. http://beagleboard.org/

Although you're making a "glass half full" kind of prediction, it's not hard to imagine that the opposite of your guess might occur in the US: All the other ARM licensees might see this as a fantastic coup for Broadcom, and follow suit with their own competing $25 - $35 boards.

After all, Texas Instruments already has their own $5 SoC available and used in their BeagleBone, so they could quite easily remove features from that board and release something into the Raspberry Pi price niche for education. (The BeagleBone's $89 places it far outside the Raspberry Pi's price niche.)

The Chinese will of course follow suit with boards based on their wildly successful Allwinner A10 ARM device, which is far better than Broadcom's SoC (on specs) and only costs $7 in production volumes. Expect a pile of competitors from that quarter!

http://beagleboard.org/static/beaglebone/latest/README.htm

it wasn't the standards body that "called them out", but the retailers who asked "shouldn't it have...."

The Pi devs thought it didn't need the CE mark because it is an unfinished product, not a consumer device (eg it doesn't come with a case). they thought this because the Beagleboard is a similarly 'unfinished' product and it too doesn't have the CE mark.

The Pi people are going through the CE motions to make sure they're covered, and finding out if they really have to go through the compliance checks on the side.

ArsTechnica does a much better job describing the issue.

There are lots of little single-board computers at low price points. It's not just Arduno and Basic Stamp machines any more. The BeagleBone, at $89, is available now. Runs Linux on an ARM chip.

The Raspberry PI $25 price is vaporware until they ship in quantity. Remember OLPC. They never made their $100 price point.

> There is a buzz around it - they aren't selling at a profit really.

Yes, yes they are. They have explicitly said that they are assembling in China so that they make a margin on each unit sold for future R&D. Just because they are a charity does not mean that they do not want to make a profit.

Anyway, once a school factors-in a TV to be connected to each Pi then it's not such a bargain.

Other projects, as requested: Bifferboard and Beagleboard. Both longer-established and actually in production. Bifferboard costs the same as Raspberry Pi Model B.

There are no holes in the board for mounting it to a case which seems like a major oversight. Maybe the RPi is good for education, their mission, but for my own projects I'll probably go for a BeagleBone. It costs about three times as much (the RPi is absurdly cheap), but at least it has documentation and mounting holes to go with its 50% faster processor.

There is a possibility that a design very similar to the BeagleBone will be spun with this new Embedded Open Modular Architecture/PCMCIA standard as well as other ARM soc designs. It will probably be at much lower cost than the BeagleBone $89 USD. http://beagleboard.org/bone

This new standard allows you to plug in whatever cpu module you wish that is compliant.

If one is in a hacky sort of mood I would suggest
A Beagle Bone http://beagleboard.org/bone
and then add this http://redpinesignals.com/Products/Chipsets/RS9116.html
and maybe this http://search.digikey.com/us/en/products/ENC424J600T-I%2FPT/ENC424J600T-I%2FPT-ND/2126005
It is only 100Base-t but it would probably be good enough for the connection to your broadband.
Linux plus less power use. Of course a LOT more work.

For that matter, are any available now?

http://www.embeddedarm.com/ aka technologic systems

http://beagleboard.org/

http://www.friendlyarm.net/

Look in the embedded market, not the FPS gamer enthusiast market.

If you are looking for a small mobo with Linux perhaps your best choice woud be the Beagle board.

For lower capabilities, Arduino would be the obvious choice, it's programmed in C, using gcc.

I don't see too much in this Jumentum, offering a web server in a chip is interesting, but this capability has been available in small chipsets (not single chips) for Atmel or Microchip PICs for years. If I needed that capability right now I'd probably go for an Arduino with ethernet.

Apart from this, Jumentum is a poor name choice, "jumento" means donkey in Portuguese.

Beagleboards are 149.00USD and Pandaboards are 179.00USD you then just need an SD card 4G or better. I run a pandaboard myself for some D-Star ham radio stuff.

A heck of a lot cheaper than the Gumstix board I bought a few years ago! and about 1/6 the price of the Beagle Board I bought last year. Looks interesting.

Yeah, this is much more like a Buglabs BUG or Leaf Labs Maple. It's also close to the price of the BeagleBoard, which is much more capable.

Uzebox is cool!

While on the subject, I'd also like to point out some other projects I've found interesting:

OpenPandora, a community-designed, Linux-running handheld. The specs are pretty impressive, by today's standards, but were even more impressive when it was first introduced. Best thing is, they're now manufacturing and shipping!

For those who like to tinker themselves, there is the BeagleBoard, a cheap (as they come) single board computer with impressive specs, designed for open source software. The Wikipedia article lists a number of alternatives, some of which may be more powerful and/or cheaper.

One interesting alternative to the BeagleBoard is the Hawkboard, which is backed by its own community. It's slightly less powerful than the BeagleBoard, but, at 89 USD, also costs quite a bit less.

And then there's the ever-popular Arduino, which comes in several varieties. You can buy them assembled starting at about 20 USD, or build your own for under 10 USD. They can be extended with "shields", e.g. to get extra I/O capabilities. Pretty cool stuff!

Personally, I am still tinkering around with resistors and transistors and the like, designing and simulating circuits with Qucs (which I feel is a lot more production-ready than that website suggests) and my Nokia N900, but any of the above hardware looks like it might be a nice next step up.

Kinda looks like a mini version of the Beagle Board, fills the gap between 8-bit Arduino and the powerful (but >$100) Beagle.

Also check out BeagleBoard. It's an ARM board.

http://beagleboard.org/

Try this one.

Sounds almost like a BeagleBoard, though that might be overkill compared to what you described.

Maybe I've just been lucky with my SheevaPlug.
I installed debian + squeezeserver on it; plugged in a 1Tb drive and I haven't need to mess with it since. That was about two years ago (OK, I do update it occasionally, but as it's firewalled off The Net and just plain works; I'm relucant to mess too much :)
Absolutely wonderful little box, and couldn't be happier with my Plug

Two USB slots, eSATA & WLAN are all nice additions. Bur I'm still struggling to see the point of bluetooth (it's mostly going to be used as a headless server -- not much point for mice/keyboards; can't see the point of an A2DP headset connection) -- Plus, any usecase that could do with Bluetooth is probably better served by the Beagleboard
Maybe I'm just not trying hard enough :)

Really well here you go.
http://beagleboard.org/hardware
http://gumstix.com/
There are a lot more but beagleboard is the closest I have seen to a mini ITX board.
Just plug in a keyboard, mouse and monitor and you are good to go.

For OEMs this is interesting but so dangerous.
Until they get some big product to use it they are very risky for a small company to buy.
You could design your product around the part only to be SOL when they go out of business. Molds are expensive.
I would so like to use one of these but until they are in mainstream production they are just too dangerous.
However for some one with a hacker spirit might I suggest that you combine this screen with this http://beagleboard.org/hardware-xM
And go to town. Maybe find an old notebook and take the case and keyboard?
Or you could make yourself a Car computer or goodness knows what else.

I don't know about mini ITX, but the Beagle Board provides a basic and fairly inexpensive option for experimenting with an ARM Cortex A8.

If you can't trust the client, a VM is of limited use(not zero use, the union of "the set of machines with malicious Browser Helper Objects that steal banking credentials" and "the set of machines with keyloggers" is almost certainly larger than "the set of machines with keyloggers"); but once a home user box is 0wned, there is very little stopping malware#1 from inviting malwares#2-#N as the situation dictates.

At some point, at least for banks and accounts with real money in them, it will become economic to ship dedicated appliances and skip the LiveCD/reboot/hardware incompatible/etc problem entirely. There are several possibilities: Imagine, for instance, something like the Beagleboard, but stripped down(no need for that fancy CPU or most of the I/O, something cheaper can load the bank website), and locked down: sealed in a tamper evident plastic box, CPU has on die verification of the bootloader, bootloader will only load signed system image, etc. All that tivoization stuff that gets the Trusted Computing Group excited. Should be under $100, possibly even under $50, in reasonable volume and nigh impossible to crack by software means(and hard to crack by hardware means without the target noticing. It doesn't really matter much if some hobbyist manages to crack his own, with prolonged physical access, that is his business). Just plug in a monitor, ethernet cable, keyboard, and mouse, and away you go.

For the terminally clueless(no pun intended), for whom peripheral hookup is a bit daunting, there would be nothing stopping you from charging a touch more and shipping a whole netbook. Even full x86 netbooks can be found at ~$200 with fair frequency, and nasty little PDA-in-a-netbook's-body offerings have been under $100 for a while now.

If even networking is too much of a challenge, you could go the Amazon route of baking in cell access: with proper caching and/or the use of a dedicated application preloaded on the client, the amount of data transfer for most people's banking needs would be tiny(and banks love adding monthly fees, so I'm sure they could find some way to recover the cost).

Android runs fine on Arm and something like the Beagle Board can decode HD fine, has lower power consumption and costs less per chip. I wonder why they would choose to go with a higher cost higher power consumption chip.

Or, put differently, if YouTube and Hulu gave users a choice between h.264 and Theora, everyone (except the freetards {...}) would choose h.264.

Yes, and ? Your point being ?
At least, we freetards would get something to run on opensource such as Firefox, on other Theora-browsers such as Opera, and on our hobbyist and community projects. One bad solution (even more if it is only bad in 10% of situations in fact) is better than no solution at all.

Freetards running Firefox, Opera or other Theora supporting web-browser (F/LOSS version of Chromium) will have something, in a dual H.264/Theora world. In a h264-only world, they would be forced to switch to the binary Google Chrome and Internet Explorer, or go back to using BLOBs such as Flash or system codecs.

Say, I'm a freetard. Say that I happen to work on a community project. Like developing the OpenPandora, the TouchBook, or the Beagleboard - well just about anything which is not a x86 device.
Such project use rather custom hardware for which no software exist. Using open-source solution is the only way to go. (these examples use derivative of Angstrom Linux).

In a world were h264 is the only solution, users in jurisdictions where MPEG-LA's patent can be enforced are left in the mud.
Using ffmpeg would be considered illegal (and a community project might be big enough to attract the wrath of a patent troll). Packing Flash or some other binary software is not an option, for lack of support from vendors.

In a world were Theora is also available, the users of such device would be happy to at least have this, even if in 10% of situation the quality is worse.

By luck, the 3 devices I mentioned have hardware for decoding h264 inside their OMAPs, so they won't probably suffer from this problem (is there a VA-API or whatever released to that chip ?).
Other community project might not have this chance.

You are right in many ways.
There is a shortage of Arm Boards but there is this one. http://beagleboard.org/ which will work just great if you don't need an HD or are willing to live with a USB HD, flashdrve , or have a NAS at home.

The new Tegra 2 is really looking good but I have not seen any small board for it yet. I am hoping that we will see a BeagleboardII when the OMAP 4 comes out. Maybe they will throw on some SATA connectors and more USB ports on it.
The one problem I see is the lack of Flash for it. That may change very soon.
I think that a Beagleboard with a good sized flash drive would make an almost viable desktop replacement depending if you can live without Flash.

If you could just get Flash for the Beagleboard I could see this as a modern Commodore 64. Right down to the external drives.

The beagleboard. Capable of dvi and hdmi output. 256M ram. Linux running on it.

At the moment, available arm processors are still behind the atom in performance by a fairly large margin, and ahead in power consumption by a similar margin. The current top of the line arm chip is the cortex-a8 used in the beagle board and gumstix systems-on-a-chip. When dual core and quad core arm cortex-a9 processors become available, that might change.

We are currently in the "roll your own" stage of development for arm machines.

Buy a beagle board or a gumstix, attach it to an lcd, mini keyboard and battery, now install one of the handful of linux operating systems available for it and you have an arm netbook.

http://beagleboard.org/

There's little incentive for the traditional desktop makers to deliver cheap thin clients. It further erodes their already thin margins for desktops, reduces revenues, and puts a whole support industry out of work.

There are a handful of companies making low-end computing devices based on highly integrated chipsets with some processing power. Freescale has some that aren't bad, and TI's OMAP 3530 series (see http://beagleboard.org/) is a good candidate. The definition of thin client will need to change, too - it'll become a diskless device that can run a virtual desktop off a server *or* a centrally managed browser using web-based apps (where rendering, playback is local.)

You also need good amount of CPU and RAM. Websites and especially video streaming and flash games are quite heavy, and so are the heavily-ajaxied Google apps.

To be precise : Adobe's crappy flash plugin and most current Javascript *interpreters* require lots of processing power.
What I mean is that the current slowness of Flash and AJAX is mostly due to current software not being efficient enough.

But google has quite some budget to leverage - both in terms of cash and brain power - to tackle this problem.
Their effort on Chrome's (browser) flash JIT compiler show that this is actually their intent.
(Similar as other work in other browsers. It seems that every browser conceptor is currently trying to make Javascript less CPU-power hungry).

Now if Google also could use their resource to bring a decent open source Flash plugin that isn't a huge useless junk (For example: finishing to make Gnash compatible and making it efficient ?)

Nonetheless, there are current (closed, proprietary) implementations of flash already running on embed hardware, so it should be achievable by google.

Last but not least : CPU performance of ultra low power embed CPU is currently rising - The next generation of ARM Cortex A9 is supposed to provide dual cores for the same power envelope as current single core Cortex A8.
In addition to that, handheld and palmtop CPUs usually have some special purpose hardware in addition to the ARM CPU (usually some DSP/FPU and some PowerVR 2D/3Dcore) - so the most CPU intensive task - decompressing the video streams - could be done in hardware.

And like someone said, shown hardware had 32 GB SSD card, which isn't really dirt cheap either.

That's probably because it is the smallest SATA SSD that you can quickly buy nowadays. They just went for a quickly customised Netbook using off-the-shelf parts.

But keep in mind that Chrome OS is basically just a browser-as-a-GUI running over a simple graphic server on a linux kernel. A minimalist Linux distribution is pretty much enough.

I've personnally already managed to cram Linux installations on 4GB Compact Flash modules (you can plug them directly into a IDE connector given the proper cable. CF and 16bits PC-CARD are basically ATAPI with a miniature connector.)
You can find projects like Damn Small Linux which pack much more functionality on minimalistic LiveCDs. (On 50MB mini CD !)
As another example OpenMoko manages to cram quite a few linux tools into 256MiBytes images.

We're really far from the minimal 16GB requirement of Windows Vista and 7.

In theory you could run Chrome OS into something like a Pandora with a bigger screen. And could indeed jury-rig something like this using Beagle boards.
That means having a Chrome OS low-power machine build *today* out of *hobbyist* parts. Now think about the near future, with mass produced units.
By 5-10 years, as the GP wrote, it's entirely possible that you could find such hardware with a bigger screen and a slightly better CPU within reasonable costs.
Even earlier than that I think.

The BeagleBoard draws a maximum of 375mA when powered from 5V. This is the whole system running at full tilt, with an SD card etc. That equates to a power draw of 1.875W (0.375 x 5) and realistically you're going to be looking at a much lower power draw than this in regular usage.

I have a BeagleBoard with Ubuntu installed and did an apt-get ubuntu-netbook-remix on it. It took a few hours of pretty much 100% CPU utilisation and the chip was barely warm to the touch...

Power figures are quoted from the latest Hardware Reference Manual - warning PDF link...

At the time Acorn simply used ARM to compete with Intel chips, in 1995 when the StrongARM Risc PC came out it was 233MHz, where as the latest Intel Pentium was 200Mhz or so.

Actually...

The ARM largely predates the Intel hegemony. Acorn designed it in about 1984 as a successor to the 6502 for the simple reason that they couldn't find any other processor that was fast enough to compete with the 6502! State of the art for 16 or 32 bit processors then were chips like the 68000, which had lousy interrupt performance.

The first actual ARM computer was the Archimedes, which shipped in 1987. (Prior to that it had existed only as a second processor addon for the 6502-based BBC Micro.) It used a 16MHz ARM2. At the time the only Intel 32-bit processor was the iAPX; the 286 ruled the desktop market. The 386 would come out in 1988.

As you say, the RISC PC came out in the mid-90s with a StrongARM, but by then Intel was working on the early Pentiums, and had pretty much won the desktop market.

The fastest ARM processors today are only 806mhz...

I have a SheevaPlug running at 1.2GHz; actual speed for integer stuff appears to roughly equivalent to a 500MHz Pentium. (Benchmarks here.)

I'm just here hoping somebody ports Risc OS Open to x86, Apple managed it after all.

Impossible, alas. RISC OS is a huge pile of undocumented hand-written ARM machine code. You'd need to rewrite it from scratch, from the ground up, to change architecture. Even just cleaning up the worst bugs (friends don't let friends allocate memory inside interrupt handlers) looks to be infeasible.

OTOH if you want bug-for-bug compatibility, someone has it working on the BeagleBoard...

I've been a fan of the ARM for years, ever since I encountered them in high school in Acorn Archimedes computers. The instruction set was so elegant compared to the i486 and Motorola 68k series chips that it was up against at the time. Flat memory model, none of this segment:offset stuff on the intel platform and a really well-thought-out streamlined set of core instructions.

I've recently got my hands on an ARM platform, and compared to what I was playing with in school, this thing is light-years ahead. 600HMz ARM, 256MB RAM, 256MB NAND Flash, GPU with ~10M polys/sec, SD Card Interface, High-speed USB 2.0 etc etc. It's all on a board that's 3" square, draws something like 1.75W at full tilt (it is powered from one of it's USB ports) and costs $150USD. No moving parts, not even a fan. 100% solid state.

I'm currently running Ubuntu on it, but there are other systems like Angstrom and QNX that will happily boot on it as well. Boot the OS off SD card, swap them out to switch operating environments and it's all good.

http://automatica.com.au/blog/2009/10/the-beagleboard/

http://beagleboard.org/

I've got no affiliation with Texas Instruments or anything like that, I'm just a happy customer who is amazed at the power of this platform, it's low cost, low power usage and flexibility opens the doors to doing so many things with it...

http://beagleboard.org/
http://www.gumstix.net/