For several years now, the need to get RISC OS running on faster hardware has been pressing. It may not be the platform's worst headache (lack of developers is surely the biggest problem), but the Iyonix and A9 hardware are now well past their prime. Even on release they weren't blisteringly quick, and cheap PCs have now left them far, far behind. Things that are taken for granted on a Windows or Mac OS machine, such as watching video, are simply impossible on current generation RISC OS hardware.
 

The options

There's more than one way of fixing this. Already many people (myself included) use RISC OS on an emulator. Whether you use the commercial VRPC or the free RPCEmu, this is the cheapest way to get reasonably fast hardware. However, many people would prefer to use a 'real' computer. For a long time this has looked impossible: Castle have stopped producing Iyonixes, and the silence from Advantage6 has been deafening for ages. However, now that the Castle branch of the OS has been opened up, the possibility of porting it to some of the more up-to-date hardware out there has become possible.

The challenges

Of course, saying it is easier isn't the same as doing it. Just because RISC OS 5 is capable of being run on new ARM-based hardware doesn't mean it's easy to get it working. It's fair to say that parts of the OS were written in a less than optimal fashion. Even with Castle's HAL layer to aid portability, the long periods of stagnation in OS development make porting the whole system a daunting job.
 
In addition, ARM hardware is unlikely to match the speeds of cheap Intel boxes for some time (if ever). There's next to no chance of RISC OS running natively on a standard desktop machine as powerful as a home or office PC. Instead, a more sensible target is the netbook - low powered, lightweight machines designed to perform tasks such as web browsing and running office applications. If that sounds a bit disappointing, it shouldn't be. Most of the ARM-powered netbooks planned for release in the near future will be much faster than the Iyonix for most tasks. What's more, there are plenty to choose from. The ARM's traditional strengths of low power play well in such a market, and companies are rushing to release machines with ARM cores to market.

The OMAP Project

Unlike Intel processors, ARM chips aren't actually made by ARM. The UK-based company licenses technology to other chipmakers, who then produce ARM-compatible hardware. One such company is Texas Instruments, who market the OMAP range of processors. There is a bewildering array of devices with the OMAP label, many of them designed for embedded or industrial uses. However, the catchily-named OMAP3530 is a bit different. It's been used to power something called the Beagle Board, which includes all the necessary gadgets to power an external display, handle USB devices, make use of on-board memory, and so on. In other words, it's basically a computer.

No one would accuse the Beagle Board of being a power-hungry monster. It only draws 2W of power, and has some key limitations (at the time of writing, an upper limit of 256MB RAM and 1280x1024 max resolution are probably the most important). But these are offset by some intriguing possibilities. The core processor is very fast by RISC OS standards, and the board also has goodies such as a Digital Signal Processor, which could potentially make video playback a reality. So it's a great a target for a port, and that's why it's been the focus of the recent work at ROOL.

Getting it working

The Iconbar's own Jeffrey Lee has been the driving force behind the project. Over the past few months, a branch of RISC OS 5 has been steadily gaining more and more functionality on the diminutive board. It's now possible to boot up into the Desktop on both the two main revisions of the Beagle Board (including the older Revision B). This has been demonstrated at a number of RISC OS shows recently: there's a screenshot here.
 
The Beagle Board doesn't have a CD or floppy drive connected to it. However, it does have a built-in SD card slot. In order to get RISC OS 5 working on the machine, you need to download the RISC OS 5 OMAP build from the ROOL website, copy this on to your SD Card and add a boot script to get it all working. This isn't really a job for the non-technical user - the Beagle Board is a development board, and is designed to be used by people who know their way around hardware. However, for those with the necessary skills to finesse the problems, there's more detail on this process here.

Progress and prospects

So, is that it? Do we have the replacement for the Acorn- and Castle-era hardware we need? Well, not quite. The port is still semi-complete, and there's a lot of work that needs to be done. At the time of writing, this is the state of play:

• RISC OS will boot into the desktop on a Beagle Board. You can use a copy of the ROOL !Boot sequence plus desktop applications (called a disk image) to gain a workable system.
• You can also use a keyboard, mouse and USB storage devices to load and save files. The drivers for these are currently the focus of a lot of development work. Over the past few weeks, the speed and reliability of these devices has been greatly improved. Even networking is now possible through the use of EtherUSB.
• There are some key limitations. There's no video acceleration enabled, and no audio at all. You can't increase the amount of RAM, and the CMOS settings your computer needs to remember your preferences across sessions aren't stored.
• Of course, this being a work in progress, there are bugs. Some are known about, other will no doubt emerge as the system becomes more complete.

What Jeffrey and ROOL team have done is show that RISC OS can run on a Beagle Board. This is a massive achievement, and one that potentially means the OS can be run on proper hardware based on that technology (such as the Pandora games machine and the TouchBook netbook).

Chipping in

Over the next few month, we can expect to see further improvements. The OMAP port isn't exactly polished and user-friendly yet (there are still components that have to be built by hand, such as networking), but it's far more capable than many would have thought possible a year ago.
 
What this project really needs is support. Although progress has been impressive, it would go faster if more people were helping out. Here are some ways you could get involved:

• If you're feeling adventurous, get hold of a Beagle Board and do some testing. The ROOL team and Jeffrey need to get hold of detailed bug reports for both the OS and its associated applications if they're going to be able to make the port complete and stable.
• If you're a C or ARM assembler developer, why not contribute to the porting effort? The ROOL site has plenty of documentation to get you started. You don't even need a Beagle Board yourself - much of the recent work on the USB drivers is relevant to Iyonix users, for example. There are also other projects, such as getting RISC OS to build with the GCCSDK that would make a real difference. Have a look here first to see what's been done, and then take a look in the forums for further help.

If you are interested in getting hold of a Beagle Board, why not ask on the ROOL forums for help? They're currently available from Digi-Key, but, as mentioned above, they come as a bare board with no case or user-friendly peripherals. With any luck, though, these diminutive kits will prove the catalyst to preserving RISC OS as a hobby OS for the future.

Links

ROOL website - Hosts the Shared Source version of RISC OS
Beagle Board - The current focus of porting efforts