The Vision of ART  -  RISC User Interview

by Alan Bonsor

Alan Bonsor interviews Peter Bondar, Managing Director of Acorn Risc Technologies, about the future of Acorn and RISC OS.

What is your opinion of some of the bad publicity that Acorn received in the press prior to and following the announcement of the new Acorn/Apple joint venture, in particular the "This Acorn has gone to seed" article which appeared in The Sunday Times around the time of the announcement?

Quite positive, actually. The level of vitriol on the Internet as a result of that article was such that, even if we completely screwed up everything else for the next six months, people would still buy our equipment out of spite!

Except that Acorn Internet users represent a very small proportion of the entire Acorn market, and their response does nothing to counter the damage done to Acorn's image with the general public.

Agreed. However, we asked a few semi-independent people for their opinions of the article, and it was clear that it was such an emotional article that any rational person would see it for what it was. As far as I'm concerned, I've seen infinitely more damaging articles about Acorn than that one. We actually faxed it to a person who was purchasing some of our technology as a positive thing to indicate just how blinkered a viewpoint some journalists have of Acorn.

Xemplar has stated that it will be working on improving interoperability between RISC OS and Mac OS. What is this actually going to mean? Are we going to see better file compatibility? RISC OS can already support most graphics and sound formats through third party utilities. Is it going to be better networking support; Acorns running on an Appletalk network, for example? Where are the improvements actually going to be made?

We are working on two sets of activities. The first involves networking and the second involves media and object compatibility. As far as networking is concerned, we plan to produce a version of Access which will allow a RISC OS computer to sit in a peer-to-peer network alongside an Apple computer. In the case of OmniClient, we plan to add Appletalk support, which we had previously stated we were going to do anyway. This will allow a RISC OS machine to sit on an Apple server in a client-server type network.

With regard to media and object compatibility, the SkyNet project which delivered PC Exchange last October will now, for want of a better expression, do Mac Exchange. This will perform a similar function to PC Exchange; you will be able to take a variety of Mac objects (sound, video, text, documents and graphics) and convert them into something which can be read by RISC OS applications, and vice versa.

Is this data exchange and conversion going to be done using existing software technologies, building on applications such as MovieFS, ImageFS, etc., or will it be developed in-house from scratch?

No. It will probably take the best of what's around and add to it; so from an intellectual viewpoint it's not a particularly demanding task.

I've heard on the grapevine that a new version of RISC OS is ready to ship. Is this true or is it just rumour, and what version number will it have?

We do have a new version of RISC OS ready and waiting for the StrongARM and ARM810 processors. We actually received our first batch of StrongARM chips yesterday. The StrongARM co-processor board has already been designed and can now go to the manufacturers, and we hope to be able to demonstrate RISC OS running on it within a few weeks. The version number has yet to be decided.

Does this version of RISC OS offer new features and tools or is it simply a 'patch' upgrade to allow it to run on the new processor cards?

It is a major rewrite which has taken several programmers almost a year to develop. However, as far as new features are concerned, there aren't any. It just runs very fast.

I believe that the ARM810 and StrongARM cards will be launched this summer. Presumably this version of RISC OS will be launched at the same time?

We haven't yet decided on the exact packaging of the various products. Because ART now has its own customers, the enthusiasts, VARs, international customers, etc., we may well choose to do things differently from the education joint venture. Most of our software technologies can only benefit from the extra power of the processor. It's debatable whether the education market requires all the power of the StrongARM, but we know from a technological perspective that no matter how much power you've got, you can never have enough. We can currently run four copies of ARPlayer simultaneously on a Risc PC700; somebody out there will want to run nine copies, or will want to play them bigger or in more colours, or whatever. The StrongARM is therefore absolutely fundamental to our technical and product strategy, and we'll be doing everything we can to encourage everybody to use StrongARM. Returning to your question, it's likely that the next version of RISC OS we release will only be supplied with the StrongARM and ARM810 processor cards.

If this version doesn't offer anything new, then, are we ever going to see an enhanced version of RISC OS incorporating lots of flashy new features? Clan disc 5, for example, demonstrates features such as Mac-style renaming of files and directories. Will such features appear in a future version of RISC OS?

The major thing from that angle is the Oracle network computer project that we're involved in. There is a number of substantial developments taking place for that project, and it's our intention to take whatever is appropriate from the project and apply them back to RISC OS. For commercial reasons I cannot discuss what these are, because negotiations are still taking place on access to certain technologies. However, the fact of the matter is that RISC OS is at the centre of a project which is getting funding an order of magnitude greater than our past investment in RISC OS.

You appear to be very enthusiastic about the future of RISC OS. I have heard that Acorn's board, on the other hand, predicts a very steep decline in the sale of RISC OS machines. Who is now actually responsible for the future of RISC OS, and who is setting its direction?

Given the level of losses that we were incurring last year, a simple management viewpoint said the we simply couldn't carry on doing what we were doing. Now it's a bit like steering a ship: you start to apply a corrective action, something else happens, and as a result of that you appear to be over-correcting, so you start to compensate for that. Look where we were with Acorn and RISC OS when David Lee inherited the situation last year. He decided to apply a lot of corrective action and then; whether we were lucky, or whether it was thanks to the people involved or whatever, we started to get things back on an even keel. The 'RISC OS is dead' attitude which was probably at its zenith a couple of months ago was almost an overshoot from this point because by then the Oracle project had started to gather momentum, with RISC OS being a core component.

The Oracle agreement effectively endorsed RISC OS as a practical working operating system, and as a result of that we are now inundated with people and companies who are very receptive to RISC OS in a way that would have been inconceivable a year ago. The result of this is that we have a situation where RISC OS has returned in a variety of guises to being a mainstream product for the Acorn group. This is one of the reasons that we changed the name of ART from Applied Risc Technologies back to Acorn
Risc Technologies.

Today, rather than having to explain about a rather quaint little operating system, designed in England and used by a couple of kids in education, we have an operating system that is liked, paradoxically, for all the reasons that Microsoft Windows is not: it's got a small memory footprint, it's fast, it's ROM-able, etc. Not having a virtual memory system is seen a major virtue; RISC OS can load in seconds from ROM without any need for a hard drive: a key requirement of a network computer.

Who knows what the state of the nation will be in two or three years' time, but we certainly have no plans to abandon RISC OS, either from an ART perspective or a network computer perspective.

RISC OS is just one of a multitude of technologies which ART licenses and sells to its customers. I believe that you licensed RISC OS to a marine navigation company recently. Could you tell us something about that deal and where it's going to be used?

That's correct. Raytheon Marine Europe is part of a large American conglomerate specialising in all sorts of things such as Patriot Missiles, etc. The marine navigation organisation, which has around 3,000 people working in Portsmouth, wants to produce a new generation of high-performance, low-cost marine navigation devices. We did a sales pitch that used the ARM7500 with the JPEG and vector drawing code parts of RISC OS, and they agreed that, compared with everything else, it is the best available.

Are there any other deals like this in the pipeline?

At the moment we have approximately twenty potential customers. I recently spent a week in Japan, visiting a number of ARM licensees and other companies who are interested in our technology. Because of the positive reaction we received over there, we will be setting up a support office in Japan. We have also recently visited a number of companies in Korea who are very interested in our technology. The fundamental fact is that we have changed our whole sales and marketing attitude. In the old days you had to buy a box from us, and it was the same box whether you were purchasing it for education, home or business use. Today, assuming the deal is right, we'll sell anything to anyone, anywhere. Whether you want an operating system, source code, subroutine, schematic or whatever, you can have it; there are no restrictions.

If you have a number of companies who will be purchasing RISC OS, ARM processors, etc., and developing their own computers and in-house systems based around them, it is conceivable that there will be a substantial increase in the number of professional ARM and RISC OS programmers. Do you see any spin-offs from this feeding back into the true Acorn RISC OS Risc PC market?

Again, for commercial reasons I cannot illustrate my answer with solid examples, but you can imagine that, as a result of the network computer initiative, there are certain organisations around the world who have no desire to be left behind when the train leaves the station and therefore have to look at developing RISC OS/ARM versions of their technologies. If but one year ago we'd gone to any one of these organisations and asked them to port their software or hardware to RISC OS, the answer would have been a resounding No. In the case of one particular company, they couldn't be bothered to do it even if we paid them a million pounds; it was just too much time and effort for them. So it's undoubtedly true that as a result of all these changes, more things will come to pass than either would have happened without the Oracle deal or if we'd just carried on as we were doing.

The only problem is that because of commercial, marketing and engineering reasons, not all those things will become apparent very quickly. You can imagine that we are now talking to organisations which might be having a total change of strategy. We couldn't just announce, for example, that XYZ Corporation was about to ditch Windows in favour of RISC OS. How they present that is very much in their domain.

If there is a large number of new companies which will be developing software for a RISC OS-based platform of whatever type, one of the principal problems they face is the lack of any serious, professional development systems for RISC OS. Is it possible that new development tools are being developed for RISC OS by any of the major corporations?

It is already true that a number of tools that were not previously available for RISC OS are now being modified or adapted so that programmers can write applications on a cross-platform development basis. In fact, the first one, Macromedia Director, was announced virtually a year ago. The NewsPAD product that we have already previewed uses Macromedia Director as one of its key technologies. Oracle will also be porting its Media Objects Structure to RISC OS to run on the network computers. We may then be in a position to port this to versions of RISC OS that ART offers to its customers. In addition, technologies such as Java are also in the frame because of the network computer initiative. The reality is that RISC OS is the only horse in town that actually delivers the performance at the memory footprint at the time-to-market to satisfy network computer requirements. Lots of other companies are claiming to have suitable software and hardware, but I can assure you that when we get together and we show them our technology, and they show us their technology, in every single contest so far we have won hands down in Japan, Korea, the UK and America.

As far as the future of RISC OS is concerned, the new joint venture has publicly declared its intention to switch to Mac OS running on the CHRP platform, with RISC OS running on an optional StrongARM upgrade if desired. This sounds very impressive but also far too expensive for the current education market. From an ART perspective, as RISC OS is still (and probably will remain) the preferred operating system for education, and schools will require a low-cost computer, should ART not perhaps be using its contacts and partnerships with other major hardware manufacturers and software developers to come up with a low-cost ARM/RISC OS-based education system?

This brings me back to what I was saying about the Oracle deal pushing RISC OS to the centre of the Acorn group again. The Oracle view is that there are three strategic markets for the network computer. The first is as a corporate network computer which requires very little maintenance and solves the problem of having to install software upgrades on thousands of different systems. The second is for the home user in a set-top box configuration which plugs into a TV or monitor and telephone line, and provides them with everything from word processing to World Wide Web-surfing facilities.

The third market that Oracle has independently identified for the network computer is the education market on a worldwide basis, where cost is of primary importance. So it's kind of amusing that we've ended up with a situation where the network computer, which by default has to run RISC OS because nothing else is available, has been identified as a key educational product. It is therefore more than conceivable that organisations like Apple in the States (and as a result of that, the joint venture in this country) could end up selling RISC OS-based network computers into education.

The thought that in two or three years' time you will not be able to reach out and touch a device that essentially shows a strong parentage to RISC OS and ARM is almost inconceivable. The idea that RISC OS applets, code and technology will simply disappear is as far-fetched as Windows disappearing.

Do you think that the world is ready for a network computer of the type that everybody is talking about at the moment?

We were recently at dinner with a large American corporation who specifically came over to discuss this point. The safest view is that, if you average out all the different opinions about it, the network computer will not kill off the PC but will rather isolate it like the mini-computer isolated the mainframe, and the PC isolated the mini.

For example, in large corporate organisations, it's clear if you watch executives at work on their PCs that they are essentially in an information-rendering mode; fetching information, displaying it on a monitor, reading it and then taking certain actions such as dashing off an email, forwarding some information to someone else, etc. Most people in these corporations are not buried in complex financial spreadsheets; they are observing data and reacting to it. For these people, the network computer offers them everything they require, while at the same time relieving all the headaches associated with traditional PCs.

As far as the home user is concerned, there is a definite market for the network computer. Firstly, it can act as a 'dumb' device which plugs into the TV and telephone line and allows you to browse the Internet, send and receive letters, play games, etc., with all your data being stored on the service provider's hard drives. There are also a lot of families who already have a PC which is in constant demand at all times; the kids need to do their homework, Mum and Dad need it, etc. A lot of thought is going into how the network computer can be hooked up to the home PC to provide families with a second, low-cost PC. The main PC, with its local storage and modem, could therefore act as both a server and a gateway onto the Internet.

The third market we've already mentioned is that of education. For schools in the UK which already have a good network infrastructure, the network computer could be used in much the same way as an A3000 or an A4000. As far as the rest of the world is concerned, the Far East represents a potentially huge market for network computers. In the UK, one in four people at home have a PC. In China, the number is one in ten thousand. Reducing that even to one in two hundred represents a market for one hundred million PCs or network computers. Network computers provide governments with a low-cost way of rapidly bringing information technology to their countries.

So the jury is out on network computers. It would be ridiculous to suggest that their future is a foregone conclusion, but it's clear that for political, economic and marketing reasons, there are significant opportunities for the network computer.

For current Acorn/Risc PC users, what immediate or short-term benefits will they see from all the new technology which ART is working with? Are there going to be any major new software or hardware releases for existing users within the next six to twelve months?

In simple terms, no. Most new technologies take up to two years to filter through the various stages of development and design. The only reason we could create the network computer in six weeks was because we already had all the bits and pieces ready to go. The positive side for existing Acorn/Risc PC users is that, because of the level of interest and funding that the network computer is generating, we can now authorise projects and get on with things in a parallel fashion, whereas previously we would have had to have done things sequentially.

Unlike when ART was launched last year, you are now responsible for all enthusiast, VAR and non-education sales of Acorn-based computers and technology. You can now basically develop computers yourself, without them having to be ordered by the joint venture or anybody else. Are you going to be doing anything like that?

Yes we already are. I have four business development people who operate in the Pacific Rim, America, the UK and Europe, and they are essentially trying to sell our new technologies and designs such as the NewsPAD and Stork.

Talking about the Stork, ART's business philosophy is that you will sell to anybody who wants to buy your products, provided they order enough of them. There are many existing Acorn users who claim they would buy the Stork. What are your plans for this product?

If Brendan O'Sullivan believes that the joint venture's education customers will buy it and we feel that we have enough customers through our assorted markets, then clearly we would love to put it into production. Obviously, given the mauling that Acorn had in 1995, the last thing that I'm going to do is to build 5,000 products just to have them sitting in a warehouse gathering dust.

Is it not perhaps conceivable that you could generate orders and thus manufacture the Stork in the same way that you did the ARM7/FPA co-processor? For that product, you advertised on the Internet that you would manufacture it provided that you received a certain number of confirmed, cash-up-front, non-refundable orders. Could you not do the same with the Stork?

The same philosophy applies. The problem simply is that the minimum break-even point for the Stork is around 5,000 units.

Is it therefore likely that we'll see a Stork computer in the near future?

Because of the way the network computer initiative is taking off, anything is possible. It's just a case of setting priorities. We are actively seeking licensees for all our technologies, and we are speaking to companies in the Far East about building both Storks and NewsPADs. Clearly if one of those organisations manufactures them, it will be fairly easy for us to reimport them and, for want of a better description, hand-modify them back to the machine we would have originally built.

I believe that one of the products ART has in its catalogue is a new MPEG card. Is this correct?

We have a variety of MPEG cards in existence, and are currently in discussion with a number of people as to which ones to put into production. We have a technology called the Kiosk Box which is essentially an A7000 with MPEG card and a few other bits wrapped around it. The product has attracted a lot of attention, especially in Europe for point-of-sale, shopping and tourism purposes, so MPEG technology in the short term is a key part of our activities and interests.

 Copyright RISC User Magazine 1996
