RISC OS 4 caught on Mac OS X | RISC OS news and information from Drobe


Beta! \| About us \| Contact \| Submit news \| RSS \| Twitter	Webspace \| Tech docs \| Downloads \| BBC Micro \| Gallery \| Wallpaper

RISC OS 4 caught on Mac OS X

Published: 28th Aug 2006, 12:31:00.

ROS gate crashes swanky Apple platform

RPCEmu has been ported to the Mac OS X platform, allowing RISC OS 4 to run alongside Apple's shiny desktop. An experimental port of the open source RiscPC emulator was checked into its source code repository over the weekend after the emulator was adapted to run on Apple computers that use PowerPC processors. A build for new Apple machines shipping with Intel CPUs is also possible, although this has not been committed yet.

The port can run RISC OS 4 and 3.7 with mouse and keyboard support. The emulation of IDE hard discs and higher resolutions needs further testing, and the user interface needs more work. The emulator manages around 5 MIPS on a modest 800MHz G4 iBook - which makes the desktop fairly usable.

The source code for the port can be found in the riscos.info svn repository from the links below, in branches/ppcmacport. It is hoped when the application is tidied up, it can be merged with the mainstream RPCEmu source code. A full release is expected once the port has been completed and rolled into an Apple application package. Adept Apple users, with PowerPC and Intel powered computers, are welcome to ask for an experimental build to provide feedback.

Originally developed by Tom Walker for Windows, RPCEmu was later ported to a PocketPC PDA and to the Linux platform.

Screenshots during development

Videos of it in action
Requires recent Flash player

Video one

Video two

Video three

Links
RPCEmu and Arculator website
RPCEmu sources on riscos.info

Discussion

Viewing threaded comments | View comments unthreaded, listed by date | Skip to the end

Finally! Excellent news for which I've been waiting a long time. Now just a bit more for it to get refined and merged into the main RPCemu code, but I feel much more certain of this then VRPC ever appearing for the Mac. I kind-of understood why VirtualAcorn did not choose to release it for Linux, but surely that argument does not apply to Apple computers?

In my opinion, the Mac serves as a more logical substitute for RO users than Windows, so sometimes I find it a bit odd that the PC, or rather Windows, seems to be the obvious alternative for many RO users. Perhaps many RO users are using a Mac at the side, so I could very well be wrong here. Anyway, I'm very happy with this news! Thanks must go to Chris for managing to get it to work on his Mac!

hEgelia on 28/8/06 1:05PM [ Reply | Permalink | Report ]

So the SA RiscPC is about 30 mips. This is 5 on an iBook. I guess we can expect about 10ish on a MacBook. That's not bad as long as you aren't expecting StrongARM performance. -- Sprite

Spriteman on 28/8/06 5:43PM [ Reply | Permalink | Report ]

What happens with the mouse buttons? I thought mac just had one and Risc OS 3. Can you get a three button mouse for the mac?

Jwoody on 28/8/06 7:35PM [ Reply | Permalink | Report ]

Mac OS X supports two-button mice. You simulate the RISC OS middle mouse button using the alt-key.

diomus on 28/8/06 7:47PM [ Reply | Permalink | Report ]

To be pedantic, Mac OS X supports more than two-buttons - I've got 3 and I know of people with 5 or so, mapping some to Expose etc.

Jaffa on 28/8/06 9:08PM [ Reply | Permalink | Report ]

I'm using an Apple Mighty Mouse (Ugh!) which has 4 buttons and scroll ball.

freder on 28/8/06 11:04PM [ Reply | Permalink | Report ]

Well, this is tempting for someone who has just bought a MacBook - having finally given up waiting for a RISCStation Portable ;-)

I do find all my doubts about emulation reawakened though: Presumably one has to buy a new RO licence to run on it (remembering all the discussions about how ROM sets are only licensed to run in their original hardware).

I also get a 'what's the point' feeling. I have been a RO fan for years and it has been my main OS in all that time. Using it as the actual operating OS on my Iyonix is one thing but is it worth hobbling the performance of my MacBook with an emulator? I find Windows dreary and poorly designed so running away from that makes sense but MacOS is much better. I suppose it gets familiar and pleasant RO applications onto a mobile platform.

I do agree with hEgelia about Macs being a better substitute for RO but that just makes these questions harder to answer.

TonyStill on 29/8/06 12:04AM [ Reply | Permalink | Report ]

Messenger Pro alone is a reason for me to want emulation on my laptop, esp for dialup. then there's the usual suspects I wouldn't want to leave behind.

freder on 29/8/06 1:20AM [ Reply | Permalink | Report ]

How viable would a USB ROM carrier be?

If one were viable, that would (should?) solve all the problems of licensing.

jess on 29/8/06 10:38AM [ Reply | Permalink | Report ]

TonyStill: you could always wait for the Mac version of Virtual RPC Adjust ;@) one price.

Jess: Dongles are a clever method of copy protection, but more than likely can be undermined in the worlds of Mac and windows in a matter of seconds...

what I want to know, is when is USB support for this app / Virtual_RPC going to be availible :@D

Good work though! keep it up.

em2ac on 29/8/06 11:10AM [ Reply | Permalink | Report ]

em2ac: Not the angle I was thinking of.

Someone makes an open source design for a ROM carrier. A common interface is written (to allow other designs) The emulators use this interface to get the ROMs.

This would give honest people a means to use the emulator.

If Castle\ROL so chose, they could license ROM images to work in a software only version.

jess on 29/8/06 11:51AM [ Reply | Permalink | Report ]

jwoody: My Mac has a standard USB 3 button mouse that I used to use on my RPC with simtec USB card. Unfortunately the functions of the 2nd and 3rd buttons are approximately adjust and menu rather than the other way round.

jess on 29/8/06 12:34PM [ Reply | Permalink | Report ]

Spriteman: The SA RPC is about 200 mips, 5mips is just over the speed of an ARM 2.

em2ac: Yes you could get Virtual RPC Adjust for Mac OS, just one problem, they don't sell it. It has been demostrated at various shows, but just like the Linux version that's also been shown, it's not for sale.

flibble on 29/8/06 12:40PM [ Reply | Permalink | Report ]

This is marvellous news. I'll certainly be emailing for an experimental copy when I get home from work. I already have a two button + one scroll wheel mouse on my iMac (the Apple mouse was horrible) so I expect that to work.

Maybe this news will persuade Aaron to release the MAC version of VRPC (I hope).

cables on 29/8/06 1:19PM [ Reply | Permalink | Report ]

Well done to all involved - I was watching the development work as it happened via screen shots posted on IRC, as the graphics memory bit ordering was sorted out in each mode.

Aaron should rightly be kicking himself now for letting the opportunity to sell VRPC for Mac OS slip away by the second. But instead he seems quite happy playing silly bugger word games in the VRPC support forums, over why VRPC+Select users don't get the 8MB screen modes that VRPC+RO4 and VRPC+Adjust get.

druck on 29/8/06 1:20PM [ Reply | Permalink | Report ]

Well, apparently there were a few (minor?) issues with the Mac version of VRPC which prevented it from being released. However, I believe that was at least a year ago and nothing seems to have been done to fix them.

VirtualAcorn should fix these and release a Universal Binary version of the software as soon as possible. Only having it for Windows sucks pretty bad. They should also redo their website, which looks cheesy and offputting especially the cheap RO3 intro bit. Also, I find the Win2000 desktop on the title page cleaner and more professional looking than the cluttered 1995 RO3 desktop... RISC OS can look so much better.

hEgelia on 29/8/06 2:02PM [ Reply | Permalink | Report ]

flibble: Well spotted, that man. I must have been thinking RiscPC 600. So, kinda sucky then. =) -- Sprite

Spriteman on 29/8/06 3:00PM [ Reply | Permalink | Report ]

druck: Aaron only has the opportunity to sell products that are available. As yet I haven't delivered a commercial grade product to him so he has nothing to sell for the Mac. If you want someone to blame, blame me as the developer.

hEgelia: VirtualAcorn will try to release a version of the software for Mac as soon as possible, but due to other commitments this may not be as soon as you or I would like.

gbarnes on 29/8/06 3:52PM [ Reply | Permalink | Report ]

Is there anyway a VA5000 could be released (either free or very cheap) as a stopgap and a demo?

(With Castle's comments about open source, maybe they would allow free distribution of 3.1 ROM images now)

jess on 29/8/06 4:01PM [ Reply | Permalink | Report ]

jess: I completely second that - in fact, I'd say it should be free and available for all common platforms, complete with a special distro of RO3.1 rigged for VA5000.

I can understand if Castle is hesitant to open source only 3.1 -- it already contains pretty much all the GUI 'niceties' which continue to make RO unique today. So, perhaps a deal can be struck between them and VirtualAcorn to have a special binary-only 'promotion' version of VA5000. It could really help the platform as a whole by making RO accessible and interest potential developers + act as a 'warming up' for a possible open sourcing of RO at a later time.

hEgelia on 30/8/06 1:55PM [ Reply | Permalink | Report ]

HEgelia: Hence why I said make available for free distribution.

I agree if there was a VA5000 type package free on all major platforms (with 16M RAM and network and host FS support) it would be of benefit. First It would attract retro gamers. Then some would find the range of good software that still runs on 3.11 and may attract some to modern RISC OS. It would also help those who use RISC OS at home and something else elsewhere, and could make the difference between jumping ship or not. It may even stimulate some software developement, a free fast emulator could easily quadruple or better the number of RISC OS users.

jess on 30/8/06 5:57PM [ Reply | Permalink | Report ]

In Reply to Jess and hEgelia

At a RISCOS Ltd AGM some years back I made a suggestion to include a time limited version of VA on a PC magazine cover as a way to gain some exposure for RISCOS. I still think that this or a "free" version would help our market.

bluenose on 30/8/06 7:36PM [ Reply | Permalink | Report ]

Jess>"Is there anyway a VA5000 could be released (either free or very cheap) as a stopgap and a demo?"

I refer you to Graham Barnes reply (above). It's not a matter of having "a version" of VA to run on the Mac but to have a "commercial grade" application available to run thereon. I am sure if Graham and Aaron had the situation where VA5000 could run acceptibly on the Mac then the VARPC variant would be just as easily (and probably more sensibly) deployed on the Mac.

bluenose>I am not a big fan of emulation of RISC OS on other platforms (the image of nails and coffins keep popping into my head when I think about it....). But if you're going to allow it I am of two minds as to whether allowing something like "VA5000" (with presumably RO3.1) out there. I always thought the 77 objects per directory limit in that version of the OS was (at the very least) embarrassing - I would *not* want other platform users to think that RO was slow (emulators are) and surprising limited in the number of files they support per directory.

If the great and the good decided that RO could be "resurrected" by releasing RO on another platform (under emulation) - and the logic they used for doing this was convincing (I don't think it is though) - then I'd say release a *good* version of RO for demo not one from the time of Babbage and steam computers and 77 file limits etc., .....

AMS on 30/8/06 8:03PM [ Reply | Permalink | Report ]

In reply bluenose:

You would have to write a program to permently delete the Rom after the the trial period is up, that is if RO 4x was used. If 3.1 was used then that may not be such a problem.

MikeCarter on 30/8/06 8:04PM [ Reply | Permalink | Report ]

I'm not a fan of emulation either but my suggestion was more on how to get some one curious in RISCOS. PC magazines are purchased by lots of people and when they try out the copy it could load up a screen splash with details of how to get a real machine. Instead of evryone complaining about the state of the market lets have some thought on getting out to a wider audiance instead of lots of internal bickering about emulation, open source etc.

Yes 3.1 is limiting but you could add more info on the up to date features. Why, we could even use the Windross install system our advantage in that it could run a series of RISCOS bits just like Windross does when you install it.

bluenose on 30/8/06 9:25PM [ Reply | Permalink | Report ]

AMS: RISC OS 3.1 running on VirtualAcorn does not have the infamous 77 file limit. VirtualAcorn uses HostFS and thus supports long file names and unlimited files per directory.

That said, I fully agree that RISC OS 3.1 is not really an advertisment for our platform. It mainly has nostalgic value.

wuerthne on 30/8/06 9:30PM [ Reply | Permalink | Report ]

Wuerthner>Martin, yes indeed you're right, and I should have remembered that as many years ago I used to use VA5000.

AMS on 30/8/06 11:00PM [ Reply | Permalink | Report ]

AMS: VA5000 is not slow on a fast machine.

jess on 30/8/06 11:41PM [ Reply | Permalink | Report ]

Although being able to distribute RO 3.1 freely would be a big step, IMO putting VA5k on the CD of PC World or whatever would be negative publicity if anything. It looks very dated next to the 3D Desktops of Mac OS X or Vista.

A time-limited version of VRPC-Adjust would be good, but you'd probably want to include demo versions of some software too (Netsurf/Oregano, Techwriter, Ovation Pro, Artworks?) since people aren't going to get too excited playing around in !Draw or !Maestro (even though they are OK apps).

timephoenix on 31/8/06 1:28AM [ Reply | Permalink | Report ]

Spot on Josh, the RiscOS look is very dated compared to all other OS's even XP (the next worst looking). You can't look at RiscOS with loving anti M$ eyes but with the eyes of the public. They will think it is some kind of retro OS.

RiscOS needs a facelift.

mripley on 31/8/06 9:09AM [ Reply | Permalink | Report ]

VA5000 would be a retro OS, but it can easily be themed. It would make sense to distribute it with a few alternative icon and tool sets in application directories.

jess on 31/08/06 10:29AM [ Reply | Permalink | Report ]

Would it be possible to have a 'Live' version of an emulator - that way RISCOS higher than 3.11 could be used as no copy would be written to the host PC?

Cheers, Dave C.

DaveC on 31/08/06 2:19PM [ Reply | Permalink | Report ]

I for one hope that VA5000 shrivells up and dies along with RISC OS 3.1 itself. It's now over 12 years since RISC OS 3.5 was released and this should really be the absolute minimum version that any application supports (as a programmer I'd prefer it to be 4.0 really, but there is still a significant 3.5-3.7 userbase that it isn't too much extra work to support.)

not_ginger_matt on 31/08/06 2:30PM [ Reply | Permalink | Report ]

Surely a better approach is to decide that if an application requires a facility that is not in an older version of the OS then the user needs to install a module that provides it, if none exists then either the app or part or it does not work, rather than making a specific cut off.

That way programs that would easily run of 3.1 will, rather than artificially preventing it.

Saying the app is untestested on earlier OSes is fair. And I certainly don't think programmers of free software should jump through hoops to support old systems, but some things just work, I have tried a couple of supposedly 3.7 and above things on VA5000 and they worked fine

jess on 31/08/06 9:43PM [ Reply | Permalink | Report ]

All this provides a massive headache for developers no matter which way you try to do it, and supporting OS3.1 usually leads to compromises for later OS versions. As a quick example, Tinct uses memory from the module area for a temporary buffer to construct larger sprites when tiling big areas with small images. It also uses memory there for it's reverse [colour number]->[RGB] lookups, and a couple of other things. This memory should be in a dynamic area really, but as Tinct supports RISC OS 3.1 (for things like VA5000) which doesn't support dynamic areas this means either extra code depending on the OS version or settling for the lowest common denominator. The real kicker that has made me want to dump 3.1 support in NetSurf in true colour support. Whilst Tinct means that 3.1 machines can plot 32-bit images quite happily, the OS doesn't allow redirection to 32-bpp sprites. This means extra faffing about to do things like the page thumbnails in 8bpp then internally converting these sprites to 32-bpp. Encouraging people to use such an outdated version of RISC OS is just a headache all round. Imagine you manage to get someone new using RISC OS via VA5000. Then they download some software. Half of it doesn't work (needs extra modules) and the other half randomly crashes (requires 3.5+ but doesn't enforce it.) At this point you're going to have either: a user who thinks the OS is outdated and unstable and simply deletes VA5000; or a user who wants to try to get things working so fires off a dozen e-mails to the application authors and wastes their time.

not_ginger_matt on 01/09/06 12:01AM [ Reply | Permalink | Report ]

Thanks for the detailed answer.

On the specifics of tinct, since there are two versions, wouldn't it be best to do the big version, no compromise minimum OS 3.5 (or greater if need be) and the simpler version lowest common denominator code and 3.1 support?

With the netsurf issues, if the OS doesn't support what is needed to make thumbnails, then wouldn't a good approach to be detect whether it does (ie to allow a module to be written to provide it, if anyone is keen) and if it doesn't to prove a conventional text page?

Alternative keep the lite version 3.1 compatible and drop support on the full version.

WIth regards to encouracing users to try 3.1, if there were a free version of VA I could probably get over 20 people (possibly a lot over over) to try it. There's no way I could see a new version of the OS being given away, but 3.1 (especially if the splash screen said what an old version it was and adviertised the new) may be within the realm of possibility. I suspect a free version of VA5000 could up the userbase of RISC OS tenfold or more, the question is whether that would be a bad thing or not.

I think it would be of benefit, especially if RISCPKG had a load of free retrogames packed.

jess on 01/09/06 2:45PM [ Reply | Permalink | Report ]

The two Tinct versions are just a full version and then one without a few bits like error diffusion (ie one simply has chunks of code missing.) All this does is lower the memory usage for no extra effort. To handle memory in a better manner it'd mean developing and maintaining two separate chunks of code simply to support an operating system that has been outdated for well over a decade. I consider this as a total waste of my time. Regarding the thumbnailing, this once again results in two code paths and extra complexity. I'm not quite sure what you mean by providing a text page though, as this doesn't make any sense here at all. The issue is that all NetSurf images are 32bpp (RGBA) internally. To generate a thumbnail we need to redirect to this image as it if was the screen and plot things there. If the OS can't support screens that are 32bpp then this simply cannot work and no thumbnail can be created. We currently hack around this issue by going via 8bpp sprites, but it is a pain to do and will be removed shortly due to other changes that make it even more cumbersome. All in all, supporting RISC OS 3.1 wastes time that could be far more constructively used. These two examples are tiny in relation to the masses of other changes missing from 3.1 that are available in later operating systems, but have wasted enough of my time that I could have done a decent amount of constructive development elsewhere.

not_ginger_matt on 01/09/06 3:53PM [ Reply | Permalink | Report ]

Would you try to use System 7 til sell people to MacOS? Or Windows 3.1 to get people to use Windows? Those came out approximately at the same time as RISC OS 3.1! I'll admit that RO 3.1 is a lot closer to the modern version of RO than the other are to their counterparts but using a 15 year old version of the OS is hardly going to convince people to buy a new GBP 1,000 computer and use an OS that can't handle many everyday things!

Gulli on 01/09/06 4:16PM [ Reply | Permalink | Report ]

I meant that if thumbnailing is a pain on 3.1, if 3.1 is in use, don't provide it. Just do an ordinary text page like other browsers.

But since you have put a lot of effort into making 3.1 work, could all that be put into a 3.1 version which would be put on the download site, and that version would rarely (if ever) be updated.

(Selfish perspective is, I've just been given a laptop by work and I have VA5000)

thanks

jess on 01/09/06 4:21PM [ Reply | Permalink | Report ]

gulli: No, but then I wouldn't really want to use them at all. Using RO 3.1 is almost no difference to use RISC OS 5 in look and feel. Once it has a nice set of toolsprites and a decent desktop font set. Using VA5000 on a decent spec system is not a step down from using the same apps on an Iyonix in the same screen mode.

The step down is the programs that won't run and lack of memory and screen colours.

A fairer comparison would be using OS X 10.0 to advertise 10.5.

jess on 02/09/06 01:42AM [ Reply | Permalink | Report ]

jess: Counting the number of major upgrades from RISC OS 3.1 to today's versions would probably move you closer to OS X 10.0 than System 7, I'll admit that - but I was pointing out the age of RO 3.1. System 7 and Win 3.1 came out at around the same time as RO 3.1 and trying to use a 15 year old version to sell an OS is not a good idea. Unless of course the selling point is nostalgia. A user trying out 3.1 and then buying 4.xx or 5.xx would be really disappointed with the lack of progress in the past 15 years.

Gulli on 02/09/06 11:25AM [ Reply | Permalink | Report ]

The reason that RO 3.1 is still similar to the current versions is that they almost got it right to start with. A user buying Adjust would notice a difference though.

I think that a nostaligia approach would be good. Attract people in with retro games, Some will then discover that a fair proportion of new programs run on it. (Eg there's enough to manage a website.) They may discover Draw and how the UI is better than even the latest Mac or Windows version.

I don't think many would buy an Iyonix, but a few would go for an A9 I suspect, and I expect a fair number of VRPCs would be sold too. I think the hypothetical x86 port being discussed elsewhere would also sell well in this scenario.

jess on 02/09/06 12:39AM [ Reply | Permalink | Report ]

In reply to Jess:

Which bit of Richard's answer did you not follow? Even if you do detect the OS version and ignore useful features based on that, it's still significant work to do that for all the things that OS3.11 doesn't do.

Then there are the other, non visible problems. There are a number of bugs documented in the PRMs (and elsewhere) which must be worked around on OS3.1 but which have been fixed in OS3.5 and later. Supporting OS3.1 means that the (often messy) workarounds are still required. Features such as Dynamic Areas, which should be used in preference to the RMA when available (to prevent problems like RMA fragmentation) aren't present on OS3.1, so either you need more complex code, another layer of memory management, or just go with the lowest common denominator (and inflict RMA fragmentation on users of the latest systems). And so on, and on, and on...

Demanding that developers continue to support a 16 year old OS, just because you have a copy of VA5000 on your laptop, holds back software development. It will never be possible to use the new features in RISC OS 4 and 5 (let alone those in Select and Adjust) if we must continue to support the die-hards who insist that they must be able to run the latest software on their antique systems. Many of the new features are there to make developers' lives easier, so software development takes longer if they must continue to re-invent the wheel.

Do you actually want RISC OS software to go anywhere, or are you just interested in it as a nostalgia exercise on an emulated A5000?

stevef on 02/09/06 2:04PM [ Reply | Permalink | Report ]

Just as a point of interest - When I wrote the original Webgen application, it would run on RO2 and up. When I rewrote it to Webgen2, I deliberately used a facility that was new in 3.5. Webgen2 only works on 3.5 or later. I stopped developing or supporting the original webgen, but apart from the late great Reg Hems nobody has ever complained about webgen2 not running on an older OS version.

Dave

DS1 on 02/09/06 3:58PM [ Reply | Permalink | Report ]

stevef:

Which is why I requested rather than throw away all the effort to support 3.1 in Netsurf, it be put into a version that is left available and either unmaintained or minimally maintained, when the normal version drops 3.1 support.

My point has never been that all versions of RISC OS should be supported whatever, more that if a facility is needed that is not in older OS, and that is not available as a module or add on then that is the time that the OS should become unsupported (or have reduced functionality in the program), not purely based on age. And that should apply to all OS versions not just 3.1

But when it comes to RISC OS software going anywhere, we are now in the situation where RISC OS cannot provide what a typical user needs, on it's own.

This means two computers or one computer with an emulator. People who know the strengths of RISC OS may be prepared to run two systems or purchase an emulator. People who don't will not. This means without a new approach we are stuck with a diminishing userbase.

With a diminishing userbase, new software becomes less and less likely.

So what remains? A free version that peeple can use on their existing systems would undoubtably increase the userbase significantly. But a modern version (RPC comaptible) is not likely to be made available, but a 16 year old version might not be impossible. It would be no threat to sales of modern systems. It would almost certainly attract many former users and some new, some may even be persuaded to buy newer systems.

jess on 02/09/06 9:22PM [ Reply | Permalink | Report ]

Please login before posting a comment. Use the form on the right to do so or create a free account.

Create a new account
Forgot your password?

Search this website

This week's poll

Recent developments have left me feeling...

Assured ROS will appear on new hardware		55%
Pleased OS desktop features are being developed		10%
ROL and ROOL will eventually kiss and make up		5%
App developments are critical		10%
Dave Holden sleeps easy at night		19%

Discuss this | Archives