RISC World


David Bradforth has a few views about RISC OS... 

Over the course of the last month, there has been much debate on the RISC OS newsgroups about Castle Technology. Pete Wild, one-time director of Wild Vision and director of Castle, is trying to dispose of his company's (Pattotek Ltd) 30% share of the voting capital in Castle Technology. Pattotek hold 75 of Castle's B shares. The latest information we have shows that the remaining B shares are divided between Roy Eastwood, Trevor Honey and Andrew Beeson, who hold 25 each, while both Jack Lillingston and John Ballace hold 50 shares each. In addition Castle have 200 A shares, split evenly between Jack, John and their two wives. So in total there seem to be 450 shares (200 A shares and 250 B shares). Last time we at RISCWorld did maths 75 shares out of 450 didn't equal 30%, it's actually just over 16%.

It's been reported on Drobe that Pete has had a falling out with Jack Lillingstone and John Balance over the future of the company. Various rumours are flying about, but given that Pete made a financial investment into Castle Technology to the tune of around £30,000 he's obviously either trying for some sort of return on that investment, or, given his comments about losing faith with RISC OS, to sever his ties with Castle.

One has to wonder what is going on at Castle. A couple of years ago the company was operating out of two sites, one in Framlingham, Castle's traditional base, and one in Cambridge. Although the Cambridge site was rented, the one in Framlingham wasn't, it was owned by CTL's shareholders and directors. In addition the company had support and office staff at Framlingham and a team of developers working in Cambridge. Fast forward two years and the company has lost it's full time developers and all the support staff and is now operating out of a complex that rents offices by the week. That doesn't sound too promising does it?

Castle's high profile product, the Iyonix, isn't looking much healthier either. Firstly it uses an IOP321 Intel processor. It has already been reported that Intel wanted to dump the "failing" ARM compatible chipset business. Indeed a purchaser has already been found, the Marvell Technology Group Ltd. In their press release Marvell said that "...Intel businesses will be able to continue licensing chip designs direct from ARM Holdings PLC...", which sounds suspiciously as though Marvell have no plans to make any more chips themselves, but will allow others to, if they need to and can afford it. Although there are large numbers of IOP family processors still available, more then enough to make Iyonixs for years to come, it does seem that further development of the IOP series is looking very unlikely.

Lack of further development in the IOP series is not the only problem that could be facing the Iyonix. The other is lead. Under the European Union Directive 2002/95/EEC certain standards and limits for hazardous material in electrical and electronic equipment have been set. One of those substances deemed as harmful is lead. The effect is this, from the 1st of July 2006 it will be illegal to manufacture new products using a traditional leaded solder process unless the product is for very specific purposes. For example servers are excluded from the regulations until 2010. The problem is that the Iyonix is assembled using a leaded process. Whilst this doesn't prevent Castle selling existing stocks, it might prevent further Iyonix motherboards being manufactured without some design changes.

So why not just switch to another type of process? Well it isn't quite that simple. The new types of solder have different properties to older solder. In particular its often necessary to alter the design of any PCB slightly so that it's manufactured in a form suitable for a lead free process. Typically the new processes operate at a higher temperature than a leaded process (around 35 degrees higher according to Intel - ED), and any components being soldered, including the PCB itself need to be designed to cope with the extra heat during assembly. In addition components need to be specified that are suitable for assembly with a lead free process. This includes the IOP321 processor itself, and as far as I have been able to ascertain this is not described as a component suitable for assembly in a lead free process. Some of the Intel X-Scale processors are available in lead free versions, including the PXA270 and others in the same series, but apparently not the IOP series.

So let us try to draw these items together. Pete Wild wants to sell Pattotek Ltd's share of Castle Technology Ltd. Assuming that he wants back his original 30K investment and given that he holds 16% of the company, that would value Castle Technology Ltd at roughly £190,000. Given that the company has few assets and no full time staff it wouldn't be unreasonable to conclude that this value relates mainly to RISC OS itself. The question that needs to be asked is simple; are the rights that Castle have in RISC OS worth nearly £200,000? It depends on what a prospective purchaser would want to do with it and here is another potential pitfall. If what I have been told is correct, and I have no way of verifying it, Castle's rights to exploit RISC OS could well be somewhat limited.

What is the value of RISC OS? All that I can say is that something is worth what someone will pay for it. John Ballance has said in the past that Castle would consider selling RISC OS for the right price, but what is the right price? Some time ago Pace tried to sell RISC OS for millions of pounds, but had no takers. In the end it was sold to Castle/Tematic for an undisclosed sum. Having spoken to a number of those connected with RISC OS my belief is that this sum would have been around £250,000. So have Castle increased the value of this asset, or decreased it? RISC OS is used in a number of products. I have already covered the Iyonix and it is also used in STB/Thin client devices. The DSL4000 box was a successful product for Pace and it's quite reasonable to assume that it has been an equal success for Castle, but it's an old design and will need to be updated, and that takes investment.

Another potential problem could be the formation of RISC OS Open Ltd. At this stage nothing is known about this new company but we can, at least, take a quick look at those involved. Richard Nicoll used to run Pace in Cambridge, until he was replaced by Andrew Clifforth in 2001. Andrew Moyler was vice president of marketing for Tematic, according to press releases issued last year. Andrew Hodgkinson worked for Acorn, then Pace and then Tematic, as did Ben Avison. The company secretary is Steve Revill, who runs 7th Software, a well known RISC OS developer. So what has RISC OS Open Ltd been set up to do? Well at the present moment I don't know. What I do know is that at least one of those involved, Steve Revill, who also worked for Pace, would seem to have copies of the RISC OS sources. At least, that would be a reasonable assumption based on his posting(s) on the comp.sys.acorn.programmer newsgroup in February this year.

This must therefore raise one important question for any potential investor in Castle. Has Castle's IPR in RISC OS leaked out to third parties and if so what rights do those those parties have to that IPR?

Finally one should consider Castle's most recent accounts as filed at Companies House. These show a loss of £48,667 on the profit and loss account for the year ending September 2005. One should, of course, accept that these figures are old, but they are the most recent figures available.

We therefore have a company, Castle Technology Ltd, whose flagship product, it would seem, can no longer be manufactured after the end of June this year, who are no longer operating from premises owned by the directors, have shown a near 50K loss in their last accounts, have shed all their full time staff and who might, potentially, have a leak in their IPR. This doesn't sound like good news.

So in the end the unanswered question is this: Is Pattotek (Pete Wild's company) selling it's share in Castle because they have had a parting of the ways and Pete wants to give others the chance to help steer RISC OS development? Or is the share in Castle up for sale now because it's shortly going to be worthless?

(I feel that I ought to add a bit at the end here. I don't entirely agree with Dave's conclusions. Certainly Castle have had some difficulties. However they are still here, they might have got rid of the premises and staff for sound financial reasons. Perhaps they needed to deal with a short term cash flow problem, there is no way of knowing. Certainly there have been further releases of RISC OS 5 since the period covered by the last set of accounts, so development work is continuing. As for the Iyonix, it could be that CTL have large stocks of assembled motherboards, it could be that the machine has already ben re-designed for lead free solder, who knows? As far as I am concerned the point is this, there is the potential to grab hold of a lump of Castle and try and build some bridges. Ultimately 16% of the company isn't enough, but it's a start. Someone could do something truly wonderful if they put their minds to it. - ED).

David Bradforth