Meeting: RISCOS Ltd Preview RISC OS 6

Report by Phillip Marsden

The talk given at the May 2007 WROCC meeting was given by Paul Middleton of RISCOS Ltd. The talk was about the latest version of RISC OS, codenamed Version 6.

The room had been set up with a long table which held a number of different computers. RiscPC, RiscPC Kinetic, A9home, Windows Vista laptop core2 Duo (the latest) and probably others that I cannot remember. These machines were there to demonstrate the range of machines on which the new OS would run.

There were effectively two parts to the talk: the history and the features of the new OS. I will cover the history and hope that another member can cover the new features in a future issue.

Paul Middleton started the evening by giving the history of RISC OS from the time that Acorn licensed RISC OS Desktop to RISCOS Ltd. If one were being uncharitable, it could be said that this was a list of excuses as to why development had taken so long to get the OS to this stage. I myself found it a fascinating account and the many parts of it rang true, as I know that there are many times when things do not develop according to plan. In fact many times there is no plan. It would appear that the Acorn RISC OS is one of those times when there was no plan. A story is told that ARTHUR, the precursor to RISC OS, was said to stand for “A Risc os before THURsday”. Does this sound like a well-planned project?

RISC OS was based on the OS (such as it was) used in the BBC Micro. Paul told us that when Acorn handed over the source code, actual code from the BBC was found in the code. (Presumably he meant an ARM version of the 6502 code, but who knows?). The code was bug-ridden with work-arounds to allow for bugs. It seems that modules were developed by different people to a loose specification. Because of the lack of programming resources when a module had been completed and ready for testing, the programmer would move on to another module. Another programmer testing his own module might find that interaction with the first module was not possible because of a bug in that first module. Because there was no programmer available to rectify the bug in the first module, the programmer would then do a work-around in his module.

As Paul put it, when RISCOS Ltd tidied up the code, the bug in the first module would be found and corrected, and then this would stop the second module from working, and so it went on. A tedious, back-breaking task which cannot be quickly done. We were told that the structure of the current RISC OS is very different to that of yesterday and that future alterations can be made much more easily.

It would appear that there has been co-operation between the two branches of RISC OS (Castle and RISCOS Ltd) to ascertain the possibility of merging the two streams, but the prospects are slim, due to the very different ways in which the source code has been written. The newsgroup chatter makes it appear that relationships between Castle and RISCOS Ltd are non-existent and that there is no co-operation at all, but the account that we heard was different. That it was Paul Middleton’s view that consolidation of the two forms of RISC OS would be difficult, if not impossible, does not mean that the two sides have not spoken and attempted to resolve the technical issues.

It seems to me that a lot of hard work has gone into the improvement of RISC OS since the demise of Acorn. Hopefully this will not be wasted. Thankfully RISC OS Six can be used on the VRPC emulator so, in the event of there being no new (more powerful) ARM-based hardware to compete with the Intel hardware, the RISC OS enthusiast can still carry on using RISC OS well into the future.