I am a long term Acorn/RISC OS user, but now, many years down the line, I am in need of something our 'beloved' OS cannot provide.

To provide some insight into what I do: I am a web developer, and use constantly tools such as Adobe Photoshop, and other graphics utilities, along with a number of good and recent browsers.

Internet Explorer, although not liked by many (including me) does support many options for a developer, as does Netscape 4 and Mozilla/Netscape 6. RISC OS browsers are behind the times, slow, and above all, the companies writing them are infantile in their approach to the web.

Someone in another thread somewhere asked the question: "Are RISC OS companies badly mismanaged?" The answer, unfortunately is "Yes". RISC OS companies in whole (not in general, not for the most part, not in part, but in whole) are run by RISC OS enthusiasts who often do not have any business sense about them. They will go through loss of sales, loss of custom, loss of almost anything before they finally decide to close down. This is not a bad thing, until they complain that people aren't advertising RISC OS. Why should we? What is there in a RISC OS machine, including ease of use that is now not available on a Mac, or a PC running either Linux or Windows? Absolutely nothing. Companies then make out some garbled information about a new machine, on which the bandwagon of RISC OS supporters jumps and says "Look, look, we have something new. It is only a 1Mhz speed increase but look!"

Time to shed some light guys. RISC OS used to be good (in 1994). RISC OS hardware used to be good (up until 1996). Now, it isn't worth much more than the top of a skip. I am ashamed to be associated with such a shambles, quite honestly. I think, with the right input from the right people (and they do exist), RISC OS could be made good again. It could even be made popular. The hardware could be improved. But it will take everyone opening their eyes to what is really available to them to be able to do that. Until that day, no 1.8GHz processors, no 256MB of RAM as a minimum standard, no incredibly fast GeForce based graphics cards, no nothing. Not even ADSL.

Do we really pride ourselves on what a wonderful (cough) machine we have, or is it just that we still want to believe it is wonderful, but deep inside we know it is just an old man sitting on a park bench telling himself long stories of the war? We were there. We lost the war, but we haven't come back to haunt our attacker. We have retreated further. We have sat and stared as our once love drains away into the never never land.

Sorry to put this on everyone, but I feel strongly enough about it to warrant a post to this forum. I want to invoke the thought processes of everyone in here, not just those people who stand out above the rest as people who can't be bothered to think, but type anyway.

I don't want to start a flame war, or one that casts (too much) blame on anyone. However, I do feel strongly about a number of points, and I wonder if other people feel the same way.

I am a long term Acorn/RISC OS user, but now, many years down the line, I am in need of something our 'beloved' OS cannot provide.

You make some interesting points (which I'll come on to) but I agree - there are many things which are not provided for under RISC OS and, right now, I don't see that RISC OS offers a complete computing solution. For example I use Draw and Easiwriter for my studies (I don't trust Word and latex is like a trip back to the stone age) and I still write some code for RISC OS. However all Internet access is done via Windows. Admittedly I don't have a modem in my RS7500 but, at the end of the day, it's more convenient to have all my net based work on the laptop and there's no RISC OS laptop.

To provide some insight into what I do: I am a web developer, and use constantly tools such as Adobe Photoshop, and other graphics utilities, along with a number of good and recent browsers.

I don't think people can argue with this. There is PhotoDesk, which is very good, but it's no PhotoShop. We're also behind on web based tools.

Internet Explorer, although not liked by many (including me) does support many options for a developer, as does Netscape 4 and Mozilla/Netscape 6.

I know a lot of people don't like the webified GUI for Windows that the Active Desktop provides but IE is a powerful component. Consider that all the MSDN documentation uses it, text editors such as Editplus hook into it to provide browsing or (more importantly) hot links to documentation.

But the whole integration thing goes further. If I download a PDF file IE can open AcroReader automatically, the same for office documents. I have Easiwriter but I'm not aware that the same level of integration is possible when downloading a Word doc in a RISC OS browser.

RISC OS browsers are behind the times, slow, and above all, the companies writing them are infantile in their approach to the web.

I don't think Oregano is being written by an infantile company but it isn't really aimed at the desktop market. It'll be interesting to see Oregano 2 when (if) it appears.

As for Webster XL - well it is being written by a one man band and, as such, represents an amazing achievement. I think the problem really is that browsers in the RISC OS market are written with substantially less resources than on other platforms. In such a situation they're never going to keep up.

Why should we? What is there in a RISC OS machine, including ease of use that is now not available on a Mac, or a PC running either Linux or Windows? Absolutely nothing.

I actually think that RISC OS isn't significantly easier to use than Windows any more.

Time to shed some light guys. RISC OS used to be good (in 1994). RISC OS hardware used to be good (up until 1996). Now, it isn't worth much more than the top of a skip. I am ashamed to be associated with such a shambles, quite honestly. I think, with the right input from the right people (and they do exist), RISC OS could be made good again. It could even be made popular. The hardware could be improved. But it will take everyone opening their eyes to what is really available to them to be able to do that. Until that day, no 1.8GHz processors, no 256MB of RAM as a minimum standard, no incredibly fast GeForce based graphics cards, no nothing. Not even ADSL.

Arguably you don't need 256MB as standard on RISC OS machines - 64MB on my RS7500 seems ridiculous at times UNLESS you're doing the kind of image manipulation you are - afterall we don't have virtual memory. Also I don't think we'll see 1.8GHz processors but the ability to run on an XScale would, IMO, close the performance gap significantly.

ADSL is a problem, so it the apparent unwillingness of a certain party to at least remove the dependencies on the IOMD and VIDC chips. If we can't have 32bit only processors then at least give the hardware manufacturers the chance to really exploit the StrongARM while we still have it and move away from these scarce and expensive chips.

Sorry to put this on everyone, but I feel strongly enough about it to warrant a post to this forum. I want to invoke the thought processes of everyone in here, not just those people who stand out above the rest as people who can't be bothered to think, but type anyway.

I think your comments are very reasonable and you're not alone in the way you think. IMHO we've lost a lot of the "casual" users and are left with the hardcore who, while enthusiastic, knowledgable and willing to help, sometimes seem to be unable to accept just how bad the situation is for the few casual users left.

One thing that really, and I mean really, concerns me is the fact that the current development tools situation is a complete joke. There is no officially supported development environment - if we're lucky Pace will feel kind and give us access to the latest versions of the Norcroft C environment but if not then once dealers run out of stocks of Acorn C/C++ then that's it - not that the tools in the package will run on 32bit only processors either.

If the Norcroft tools are likely to be out of reach then why aren't RISC OS Ltd investigating the possibility of putting together a full package using GCC? They even have two IDEs to choose from now. What's missing is ResEd and the ability to compile module code.

There is no concensus as to the way forward for RISC OS development either. Is the toolbox the favoured environment and if so why aren't the ResEd tools (or equivalent) being made available? If this is not the way forward then what is?

As someone who spends most of his time now developing on linux and Windows I also have to ask:

Where is pre-emptive multitasking?
Where is multithreading?
Where is virtual memory?

People will cry out that these aren't important but sorry, certainly for the development of network applications, they make life much, much easier. If I want to multitask a network app on linux I can create a thread, block it on a socket and then let pre-emptive multitasking take care of task swapping. On RISC OS I have to delve into the Internet event, modules and manage multitasking myself.

Have the other platforms made me lazy? Yes. Are the other platforms more productive to develop on? Hell yes. Why spend 30mins writing RISC OS code when I can do the same on linux in 5mins?

The lack of these facilities also limits porting. How much effort do people think it would be to port something like Opera or Mozilla? Personally I wouldn't even contemplate it - RISC OS isn't really capable of supporting code written using certain "standard" techniques.

Where's Java? People who haven't used it won't appreciate just how fast, and easily, full scale applications can be developed in this language.

I've thought for many years that some of our problems actually stem from the fact that BASIC/ARM are built in - there has been relatively little impetuous to really develop compiler tools and libraries. I find the fact that Vantage has been developed in BASIC / ARM quite incredible. How much of it is difficult to maintain ARM code that could've easily been written in more managable C code with little compromise on performance?

I suspect people will tell me to stop moaning and do something about it? I ask why? I'm not a tool developer - I have little interest / expertise in it and they require more time than I have available. I am working on various VOTI projects but, with about 5 hrs / week currently available it's slow going. I feel that development tools are RISC OS Ltds responsibility. If they want software written for their OS then they need to provide the tools needed to make it happen.

I feel that development tools are RISC OS Ltds responsibility. If they want software written for their OS then they need to provide the tools needed to make it happen.

But RISCOS Ltd have other priorities. Please give careful consideration to the G in GCC.

If anybody is using GPL'ed code in their commercial products, they are either very thick or are being undermined by a colleague. Why, is somebody?

Pace have absolutely no reason to give up the tools and, arguably, the C compiler is actually an asset as it's certainly one of the (if not "the" best C compilers I've ever used.

Yes, much open source development is subsidised by day jobs, but so what? Much of that is created by hobbyists, and if anything is technically better as a result. Commercial companies in general write poor code. Look at Unix - most flavours are pretty awful, Linux is one of the cleanest, exactly because it was a hobbyist's obsession. Look at Microsoft. Even RISC OS went down the tubes as the commercial people took over from the original developers, and PM is having a horrendous time untangling it. That's exactly why opening up as much as possible is so important.

The disservice to the Unix folk isn't mine - it's the "mainstream" (haha) RISC OS community who don't give them the cred they deserve (yet - just wait).

If anyone is wondering why I haven't been vocal recently it is because I have become so fed up of the market that I have actually stopped bothering with RO and the irritating people contained within the market. This is the first time I've booted my RPC up for a couple of weeks, I have lost all faith and inclination and would much rather read books than p*ss away talent.

To be fair, I was a bit harsh about ROL's short-termism there, some of it is justified, and some longer-term problems are being tackled.

Dunno - I've just seen Paul Middletons letter on Drobe and, while I understand what he is saying, it doesn't inspire confidence for various reasons.

If anybody is using GPL'ed code in their commercial products, they are either very thick or are being undermined by a colleague. Why, is somebody?

Er..yes - lots of people and companies. I thought the whole point of the GPL was to allow this to happen providing improvements made to the GPL'd code was released back to the community. Doesn't mean the GPL'd product can't be released alongside proprietry, non-open source, code.

I don't follow your reasoning. What money-making capacity do Pace's tools have that would not be magnified a thousandfold by opening up the OS to new developers? Especially if they can better the GCC.

Assume for the minute that the compiler has some super-duper technology in it to optimise spilling variables into registers that beats the competition by some margin. That is intellectual property which may be licensable.

BTW I'm not saying this is the case - just giving an example that things are not always clearcut.

Yes, much open source development is subsidised by day jobs, but so what?

Simple - almost all software engineers I know have clauses in their contract which says that everything they produce, on company time or not, belongs to the company. Currently the companies turn a blind eye to the open source community because they can benefit from it but what if they decide to enforce it?

Also imagine if, one day, those same people decide that it's enough just doing the day job?

Much of that is created by hobbyists, and if anything is technically better as a result. Commercial companies in general write poor code. Look at Unix - most flavours are pretty awful, Linux is one of the cleanest, exactly because it was a hobbyist's obsession. Look at Microsoft.

I think you're over-generalising, not every company produces bad code. It's also incredibly easy to be perfectionist when there are no commercial pressures. Again this is another example of the fact that the open source model only works because it is indirectly "backed up" by commercial products.

Even RISC OS went down the tubes as the commercial people took over from the original developers, and PM is having a horrendous time untangling it. That's exactly why opening up as much as possible is so important.

What?? RISC OS has always been a commercial product.

The disservice to the Unix folk isn't mine - it's the "mainstream" (haha) RISC OS community who don't give them the cred they deserve (yet - just wait).

I don't think there is a developer (or anyone attached to developers) in the RISC OS market who doesn't appreciate the efforts and talents that available in the open source community. All I'm trying to do is point out that open source software isn't really free - it's often subsidised by industry in the same way that IE is subsidised by other departments in MS. Even software produced by hobbyists is subsidised - afterall if the hobbyist isn't working then they're being subsidised by their parents...or possibly the state.

As a friend who is heavily into FreeBSD points out the cost of developing open source software is probably just as high as developing commercial software. In addition there are training costs to consider. However the cost of obtaining the software is zero and this is what really distinguishes open source software from commercial software.

BUT - There are signs of a new mood of realism in the higher ranks (not necessarily on the newsgroups ). Pace are pushing the OS forward, RISCOS Ltd are doing good things for the GUI/desktop, Castle, MicroDigital, RiscStation et al are pushing the hardware on, and most importantly they are all learning to listen to their customers and users. Some of the good old killer apps, like Optima, are being reinvented too, and the programming community is busy playing with XML, Unix ports, SVG, file converters, web browsers, etc and generally showing signs of awakening from hibernation. Hey, we just beat Germany on their own turf again!

I personally believe that the RISC OS community still has a lot to learn from the success of the GNU/Linux model - amongst other things the current renaissance would stand a much better chance of success if the souce code was opened up. But this would/will be a long and *painful* process of social and commercial reawakening, and is unlikely to be on the agenda unless the current renaissance fails to make it. Maybe next renaissance? My latest thoughts revolve around lessons to be learned on branding and PR.

Don't lose heart, hang on in there and give yourself a Beckham mohican (well, maybe not).

[Edited by guy at 09:46, 18/9/2001]

But RISCOS Ltd have other priorities.

Like what? If RISC OS can't attract developers then it's dead in the water. No one wants an OS with little or no application development.

Please give careful consideration to the G in GCC.

I'm not quite sure what you mean by that. I have some idea but I'd rather not get it wrong

But RISCOS Ltd have other priorities.

Like what? If RISC OS can't attract developers then it's dead in the water. No one wants an OS with little or no application development.

Please give careful consideration to the G in GCC.

I'm not quite sure what you mean by that. I have some idea but I'd rather not get it wrong

Whether or not opening up the source of RISC OS itself is viable, surely the developer tools need the GPL real bad. And unlike RISC OS itself, the tools are not a great property asset - there is b*gg*r all to lose in trying.

I bet if Pace or RISCOS Ltd announced their intention to move to the GCC and a toolset released under the GPL, we would suddenly find a few more developers prepared to work on (and with) the toolset.

In fact, it needn't be Pace or ROL - just anybody with a little cred in the community. But the big names always help.

[Edited by guy at 14:24, 18/9/2001]

Like spending the investor's cash on things they can do short-term. I didn't say I agreed with their priorities

Sometimes I do wonder if anyone has long term plans for the platform.

The G is for GNU and by implication the GPL, and everything that goes with it, like intellectual freedom, technical excellence, the developer community doing the hard work for eachother and crafting the tools the way they themselves want them. Having fun. It's how the GCC found its way to us in the first place.

Yes I understand this - I was wondering if you were referring to the implications of the GPL if it was to be used as part of a commercial offering (which I also understand).

Whether or not opening up the source of RISC OS itself is viable, surely the developer tools need the GPL real bad. And unlike RISC OS itself, the tools are not a great property asset - there is b*gg*r all to lose in trying.

That depends on your viewpoint. Pace have absolutely no reason to give up the tools and, arguably, the C compiler is actually an asset as it's certainly one of the (if not "the" best C compilers I've ever used.

...and they still have to pay someone to maintain them.

I bet if Pace or RISCOS Ltd announced their intention to move to the GCC and a toolset released under the GPL, we would suddenly find a few more developers prepared to work on (and with) the toolset.

Maybe. Remember though that for the most part open source development is actually subsidised by the day jobs those developers have.

In fact, it needn't be Pace or ROL - just anybody with a little cred in the community. But the big names always help.

I think you're doing the likes of Nick Burrett and the rest of the GCC / UnixLib crew a bit of a disservice there

If anybody is using GPL'ed code in their commercial products, they are either very thick or are being undermined by a colleague. Why, is somebody?

Er..yes - lots of people and companies. I thought the whole point of the GPL was to allow this to happen providing improvements made to the GPL'd code was released back to the community. Doesn't mean the GPL'd product can't be released alongside proprietry, non-open source, code.

Assume for the minute that the compiler has some super-duper technology in it to optimise spilling variables into registers that beats the competition by some margin. That is intellectual property which may be licensable.

Yes, much open source development is subsidised by day jobs, but so what?

Simple - almost all software engineers I know have clauses in their contract which says that everything they produce, on company time or not, belongs to the company. Currently the companies turn a blind eye to the open source community because they can benefit from it but what if they decide to enforce it?

Also imagine if, one day, those same people decide that it's enough just doing the day job?

I think you're over-generalising, not every company produces bad code.

It's also incredibly easy to be perfectionist when there are no commercial pressures.

Again this is another example of the fact that the open source model only works because it is indirectly "backed up" by commercial products.

Even RISC OS went down the tubes as the commercial people took over from the original developers, and PM is having a horrendous time untangling it. That's exactly why opening up as much as possible is so important.

What?? RISC OS has always been a commercial product.

I don't think there is a developer (or anyone attached to developers) in the RISC OS market who doesn't appreciate the efforts and talents that available in the open source community.

All I'm trying to do is point out that open source software isn't really free ... Even software produced by hobbyists is subsidised ... the cost of developing open source software is probably just as high as developing commercial software.

[Edited by guy at 10:56, 19/9/2001]

[Edited by guy at 11:00, 19/9/2001]

You can sell code under a commercial license if it relies on separately-available GPL code, and put both in the same box, but not if that GPL code is incorprated into the commercially-licensed bundle.

Before I reply to that what's the difference between box and bundle? Am I right in thinking you mean that

A commercial product can make use of GPL code as long as the GPL code itself does not form part of the commercial element of the product? In addition the GPL code can be distributed with the commercial code (to stop the purchaser having to hunt it down)

If so, then yes that's fair enough.

Any software incorporating GPL'ed code automatically comes under the GPL too. IIRC even M$ have had to GPL some minor product (I forget which) when this woopsie on the part of a junior codie was discovered.

Ah well then the team leader or product manager was at fault

If you know of more people who are doing this, you can ignore their commercial license and even take them to court to release their source code.

I think I'd run out of money in a court case before a any reasonably large commercial organisation did

But it's clearly peanuts unless it can be ported to a mainstream OS, in which case its functionality will be re-enginered into the GCC before anybody can make a decent profit. No, the whole thing just doesn't make commercial sense.
Far better to release it, so that others can enhance the value of your major product, the OS, by developing apps and so expanding the market.

It's not peanuts if the technology is the algorithm used rather than the implementation. A good algorithm is portable to any OS. It's the same with patents on techniques such as MP3 compression - the algorithms are the valuable part.

I realise that my argument is thin concerning the C compiler but I'm merely pointing out that Pace may have options we are not aware of.

They'd lose their brightest and most creative programmers to a more open-minded competitor. They won't do it.

Not necessarily. Most people work to live so how many people would be prepared to take a drop in their standard of living if maintaining an ideal meant taking a 30% pay cut? (the figure is not altogether unrealistic). If it was enforced on myself then I'd probably just drop out from VOTI. The gratification I get from writing code for RISC OS doesn't compare to what I'd lose.

To see the sense of this look at what happened to Strykers Run III? Keeping their jobs was more important to the team than releasing the game.

Are you volunteering for a prefrontal lobotomy? Nor are they.

No but at the end of the day I have to be able to feed myself, pay the bills, have a reasonable life and so on and so do they. Currently I work fulltime and am writing my thesis. I really wouldn't be too upset right now if I was told my RISC OS activities had to stop (although I was forward thinking enough to ensure that can't happen).

I'm merely realistic in my priorities. I suspect various (important) people in the open source world are too.

True, eg Opera, they are the exceptions that prove the point.

I think Icon Technology proves my point - Easiwriter is fantastic. It's commercial software though.

Is this bad? Don't underestimate the urge to get that code out there.

I don't but, OTOH it's relatively easy in the open source world to go "oh this isn't quite right, I'll fix it and delay the release by a few hours". In industry there is no such luxury - a deadline is a deadline.

It certainly works *better* that way. Don't we all.

I don't wish to impose my will but it *only* works that way. Without a job how do you propose the open source developers buy a PC?

In the earliest days, the RISC idea started as an academic project, and was bought up by Acorn to make it happen. The original code had more of a blue-sky brief, and the commercial 'development' followed as an afterthought.

I believe you're referring to the ARM processor rather than RISC OS itself, and even that was started out of commercial need - Acorn needed a next generation processor and there wasn't a suitable one commerically available. IIRC RISC OS was commissioned as an OS for a product (the Archimedes) when it became clear that the "unix-style" OS for the ARM chip wasn't going to be viable.

some of them have a funny way of showing it

*grins*

True, and an essential part of the open philosophy. You have to stop thinking commercial and start thinking fun/fulfilment/happiness - but you obviously can't do this until your tummy is well stuffed with herring.

Surely this just provides support for my argument

Later on, you or somebody else can make money again, as a service provider.

If there is a service to provide, not all software requires it.

A good algorithm is portable to any OS

the algorithms are the valuable part

OTOH it's relatively easy in the open source world to go "oh this isn't quite right, I'll fix it and delay the release by a few hours". In industry there is no such luxury - a deadline is a deadline.

IIRC RISC OS was commissioned as an OS for a product (the Archimedes) when it became clear that the "unix-style" OS for the ARM chip wasn't going to be viable.

Surely this just provides support for my argument

Are we going to see a new commercial Unix or Wintel C compiler then?

I doubt it but Visual Studio is still a major product for MS and they are still actively developing and selling it.

like LZW was so valuable somebody reinvented PNG?

I never said you couldn't do alternatives

I don't think this is OTOH, I think it supports my case - I'm sick of immature bugware.

So am I. Problem is that managers are often far too optimistic about their timescales and developers often make compromises they're not happy with. End result is that they get slated and yet probably are the same people churning out fantastic code in open source.

It's that old finding-something-to-disagree-on problem again

*grins*