The Death of RISC OS
M'luds, the case before you is an unusual one. For many years, a single computer platform and operating system was the main stay for educational establishments up and down the country with a many, and varied, user base.
Many things have happened over the years, and today, we have to make the decision; how do we prevent RISC OS vanishing.
Okay, that may be a bit melodramatic, but essentially, it's correct.
The question is, why, on a CD based magazine (which means you're running at least RISC OS 3.5 or are viewing this on another platform totally), should this matter — after all, we're all happy with using 3.5 aren't we. Well, aren't we?
Well, no. And if you are, shame on you. Here's why...
Acorn and RISC OS, a potted history..
We all know the history of innovative, industrious bods who single handed created the ARM processor and an OS to run on it, brought forth some of the most powerful hardware around and along the way, managed to cock things up amazingly!
This was the first Archimedes operating system. Originally the Archimedes was going to ship with RISC OS, but this wasn't ready, so Acorn cobbled together the working parts of RISC OS and wrote a new desktop (in BASIC!) and released what was effectively half an operating system (hence the name Arthur).
26 bit addressing
Why they did that, I'll never know. They had arguably the single most powerful processor going in the ARM chips, yet decided to bodge it and cause muchos problems down the line.
Lack of finalisation
Acorn were masters of getting things 90% done, but never masters of completeness. You only have to look at OS 3.0 for that. It had more bugs than Windows 3!. 3.1 was out very quickly.... All down to not going that little extra. Sure, 90% in Acorn terms is still close to 98% in Microsoft terms (stability and usability), but it's that little extra which means so much.
Lack of focus
Acorn, while they were innovative (they were one of the first to use SCSI with the Doomsday system), they were incredibly short sighted. Many people joke about Bill Gates' comments ("computers will never need more than 640K" and "the internet will never catch on"), but conveniently forget how far behind Acorn was in terms of harddrive capacity, memory, CD drives, web browser and web technologies. It wasn't until OS 3.7 that you could really use larger than 512Mb drives without partitioning. Memory upgrades have been hopeless due to the specific type of memory required and browsing under RISC OS, while getting better, is still massively behind other platforms. They just didn't see it coming!
The RiscPC, while undoubtably the finest Acorn machine ever is not very well designed. There are hold-ups on the memory buses, the IDE system is not that amazing and if you have a mk 1 motherboard, only had 8 bit sound. It was never designed for the StrongARM which really does show the problems up (well, the motherboard was never designed for the SA which can be up to 40 times faster than the ARM 710). Other awful designs included the A5000 and A7000 machines (though the A7000 is not as bad as the A5000).
The fiasco over the Pheobe demonstrated this.
Reliance on one market
Acorn machines and education had been linked together from the days of the BBC B when the government of the time instituted the "computers in schools" campaign. This gave Acorn the tie in which was a stable source of income for many years. They never really invested any money elsewhere.
NMIC (Not made in Cambridge)
If a component was not made in Cambridge by Acorn (or farmed out to an approved company), they were not interested. This cost them dearly — very dearly. IBM wanted to licence the RISC technology (including the OS) when they saw the problems they'd let themselves into over MS. Acorn would have none of it. Imagine what would have happened had they said yes...
Acorn machines go on and on and on and on and on. While this is a good thing in one respect (they last), it is a financial death trap as no-one upgrades, especially when Acorn decided that when OS 3.5 came out, software had to still run on 3.1. That single mistake probably did more damage than any other. Apple don't backward support and Microsoft certainly don't. In that way, both Apple and MS get cashflow through punters having to upgrade. Mercenary? Possibly. Financially sound? Definitely.
Reliance on custom components
While the world and his dog were enjoying cheap components (such as memory, network and video cards etc), Acorn machines relied on VIDC and IOMC as well as other custom components. When anything is custom made, that adds an expense onto the machines overall price; sure, they do a very useful job, but it is an added cost. There is also the risk that these parts will vanish.
Acorn machines have always been fast. Even back on the BBC B days, machines (such as the Oric-1) which used the 6502 were not as fast as the BBC B.
You can just add and add and add to the RPC. The ability to have a second processor to run an alien OS on is one not seen very often. Neither is a pizza oven on the top slice...
If you could find someone in post for more than a couple of months at a time, customer support ranged from blindingly good to terrible. Most of the time it was good. That said, if you wanted technical support, you couldn't ask for anything better!
Innovation and creativity
Acorn made great strides in hardware design and development (with respects to upgradability). Their designs were very creative. The A30x0 range demonstrated that. They are some of the best looking computers designed. The RPC is another such box. .
Okay, given all of these you can see that Acorn had it's problems. The two biggest were the reliance on the education market and the longevity of the machines. In 1997, we had a change of government.
The education sector had already started to buy Windows PC machines prior to this as the RPC was starting to show it's age and the number of software packages for Windows had exploded. Sure, !Draw is wonderful for ICT, but with packages such as T3 (a powerful technical word processor) and a plethora of word processing and DTP packages, it wasn't hard to see why schools moved.
There were two other reasons for the move to the PC.
While this removal of Acorn's cash flow was progressing steadily, the lads in Cambridge were developing the doomed Pheobe. At the time, it was just what the doctor ordered. Fast, able to use off the shelf parts and above all, had a new OS and would be cheap enough for those wanting to upgrade and poke their tongues once more at PC owners who had been boasting at the speed of the 486DX 100s over the SA 233. It ran into delays, funding problems, hardware faults, OS problems and any thing else you wanted to throw in. It was shown at RISC OS shows around the country to much acclaim. But, the delays were costing Acorn dearly.
Combine this with the new deal signed by the government with Microsoft and the end was in sight. Acorn shut down the desktop division, became E14 and then, well, you know the rest.
Stand up and be counted...
There were many attempts to get the Pheobe project running again by buying it from E14. Peter Bondar, formerly of Acorn lead this. Many at the time hailed Peter as being a god-send. They were very wrong. He was a consultant who described RISC OS users as either "education establishments or anoraks". The consultation fees were rather high as well (up to £500 per meeting). At one point, it looked like our little yellow pal would come back, but (if usenet is to believed) things went sour. Mainly as the document of understanding was signed in the morning then someone signed another agreement with a dutch company in the afternoon. The rows broke out and talks broke down. In the end, Bondar walked away. Off to fly his plane...
Then came RISCOS Ltd. They didn't want the hardware, just the OS. It would then be up to hardware companies to make the machines. Some of the finest RISC OS talent were employed by RISCOS Ltd to do the development from the old Acorn 3.8 OS. At Wakefield 1999 the new OS was launched. RISC OS 4 was everything you could ever want.
There were three hardware companies; Castle (who everyone knew anyway), MicroDigital (who many years prior had been licensed to produce the first Acorn clone, the Medi) and a new company, RiscStation. Castle had already signed a deal with E14 to carry on making the RPC and A7000 under licence.
Everything was looking up except for one major problem. Pricing and public relations for the OS. Schools used to get a bulk discount for new chip sets (they would be about 20% the price which they would retail at). Unfortunately, RISCOS Ltd couldn't offer such a discount. If schools wanted it, then it was £120 per chipset. .
The hardware companies didn't help either. Typically, one of the ARM 7500FE machines would cost around £550 for a base model. While the newer machines were far better than the A7000+, they were still poor in comparison to the PC market price and specification wise. There was also never a complete comparison between the old machines and the ARM 7500FE machines. Combine the cost to upgrade to an OS 4 machine from an OS 3.1 machine and the lack of compatibility information from the company behind the OS and it's not hard to see why people didn't upgrade.
Everything was otherwise shiny. We had the first RISCOS Ltd/ARM Club show and the announcement of the palmtop (RON). There were even mentions in the Guardian about the developments in hardware and OS!
Time moved on. The hardware and OS didn't. OS 4.02 (or 4.03 for the Kinetic, RiscStation or MD machines) had it's problems, but it was far better than any previous offerings. .
The next jump in the OS was Select. Incredibly badly publisised by RISCOS Ltd — it consisted of a flyer given out at Wakefield by the ARM Club (RISCOS Ltd were unable to make the show that year). It offered many bug fixes (quite a few of which had been identified during the beta testing stage of 4.02) and some substantial upgrades (such as dhcp — essential for broadband). It also made things look pretty.
Due to a number of reasons (cost, terrible publicity, not actually knowing what you'd get and the very odd "up to 3 CDs a year", some took that to read you may not get 1 CD!), the uptake of Select was not as great as anticipated. Initial estimates were for 500 subscribers in the September of it's launch (roughly 4 months from the Wakefield announcement) and 1000 by the following April. It was at roughly 900 by the following September.
Something which had come to the fore was the amount of nastiness in the developer community. An inner sanctum of developers seem to have decided that they were going to be the future and basically ridiculed and insulted anyone else who dared to start asking questions. In a small community which is trying to get itself out of a rut, the big fish in a rapidly shrinking puddle system doesn't work. Companies willing to invest in RISC OS decided not to invest after looking in on the comp.sys.acorn.programmer group. If there was this amount of infighting, they didn't want to know. One estimate has put this loss at over £1 million in one year.
The exodus was carrying on. Then, one fine day, Virtual Acorn arrived.
VirtualAcorn - the saviour of RISC OS?
Sales of new machines had dropped to very low levels. A piece of freeware which had been around for a while for Windows machines by Graeme Barnes was taken up and made into a commercial package with a licence from Pace for OS 3.1 to be shipped with it.
Once again, schools which had moved to Windows, could use software which had been left in cupboards and boy, did they snap it up!. This reduced their costs. It was a marketing opportunity many of the RISC OS companies took advantage of by re-licensing their software for using on VA. This provided an extra bonus for VA and also an income for the dormant sales of RISC OS software. .
On the hardware front, we had very little new developments. Lots of hot air, no products. The Mico was effectively being sold by APDL (who were becoming an increasingly major software force with their partnership with Dave Bradforth's ProAction label). The R7500 from RiscStation had been destroyed by one of their developers in an attempt to show how robust it was and also the 50ns memory upgrade became available. The Kinetic 300 came and went. The MD laptop was announced, then vanished after signing an agreement with RiscStation. The RiscStation laptop was announced and we waited.... and waited.... and waited.... and are still waiting....
Very little was happening in the RISC OS market. A few bits of software were updated, but nothing to write home about. Then MicroDigital announced their Omega...
Omega — the number 1
The Omega was a miracle. It would work with anything written on a RPC (so meant an instant back-catalogue was available), could take a second processor (PC or X-Scale) and if using the X-Scale, could use the ARM-Twister (a 26 bit to 32 bit hardware convertor) and become really fast. Money was taken and to date, only a selected handful of people have their machines.
In the middle of last year, Castle sent out a missive to developers about their new X-Scale hardware. A full 32 bit OS and Xscale machine. It looked like a dream come true. Until it was realised that it had a limitation. .
Software would have to be recompiled to run under the 32 bit OS. Fine if the company (or individual) was still about and was willing to shell out £100 for the updated version of Acorn C, but not all developers were prepared to do the work for a relatively small number of sales. Castle realised this and had the Aemulor software written. .
For the price of an Iyonix or Omega you can get a lot of PC (or Mac). It is little surprise that so few have been sold.
In 2003, a deal was signed between VA, RISCOS Ltd and MD to ship VARPC on a Celeron based WinXP laptop. Many have seen this as the final nail in the RiscStation laptop (and possibly even RiscStation). Was this inevitable with the lack of cash being earned by RISCOS Ltd? Perhaps. A good idea? That remains to be seen. .
We now have a dilemma.
The use and ownership of RISC OS (both machines and the OS) is at the "critical mass" point. People are jumping ship to Windows and Linux due to the level of software and development, with their RISC OS machines being used as second boxes for the kids. It may not be too late to get those on 3.1 machines to upgrade, but only if the hardware manufacturers radically drop their prices. In order to lower prices they need to sell more units, but they cannot sell more units until the price is lowered, catch 22.
£399 for the Mico is a good price for a RISC OS box, but when you can get a fully blown OS 4.02 SA machine on eBay for half of that, it's not that good. If for one month only, all of the AMSes dropped their ARM 7500FE machines to £200, a good lump of OS 3.1 users would jump. Effectively, they'd have a new machine for £100 and the new OS as well. Of course, they'd have to advertise this effectively (not just in one Acorn mag, but across the mags and on usenet as well as any other way they can).
New machines are too expensive. Okay, using a niche market bumps up the price, but not to the extent that companies are charging. If they were sub £1000 (inc. VAT) then more would take the plunge. As it is, I know that I won't. I can't justify the cost. The 4 RISC OS machines in front of me (all except one on OS 4) are fine for what I need. I can't justify £1600 on a new machine, especially as they are comparatively so lowly spec'd.
Besides the obvious price reduction for the hardware, developers have to work together.
Forget the personal attacks and the long forgotten arguments. I have noticed on one RISC OS forum that a developer who has publically now left the market being referred to in very derogative terms for trying to dispel some of the hero worship which seems to have infected the platform. I still use a number of his apps, which (if you believe some of the rubbish said about him, will destroy the machine, kill two of your children and possibly have the US claim he has weapons of mass destruction!) work better than anything under Windows, an OS I fight with daily).
We have so few developers left that we cannot afford to lose even one more. We have so few RISC OS users left that we cannot afford to lose any more. It's quite simple, we need to have more copies of RISC OS in use. If we have more users there are more people to buy software, this would mean its worth software houses updating existing products and releasing new ones. .
How do we get more users? Advertise, find a new niche market, promote RISC OS, but do something, and do it soon. Otherwise RISC OS might well be consigned to being run, (like so many of the games machines and home computers of the 1980's), under emulation with no hardware, no new software and no development.
The views expressed in the Insider column are not necessarily those of the RISC World Editor or Publisher. We allow contraversial opinions to be expressed in this column where we feel it may be to the benefit of the RISC OS community that these should be given a public forum.