Spending a Million on RISC OS

RedSq is pretty close to VRPC – it’s working here with Netsurf etc.
Relying on old patches to make the Internet work – still sometimes -get an error from TapirMail – “Socket is not connected” so not totally working.

The sound is a bit strange – so !MP3Radio a no go – the new map program works well – following the TT course around the IOM :-)
Using RO 5.19 (19 Jun 2012) – rom loaded from disc – pity there are not later versions of RO 5.21 that can be run from disc – instead of soft loading – can make use of the separate CMOS files – so modules can be unplugged for testing.
Runs RO 3.1 or 4.02 or 5.19 from disc.
Have to softload 5.21.
Max screen memory varies – depending of the setting of VRAM – 0 gives 32768000 bytes with 5.19. (5.21 needs 8M VRAM)

Have tried someone’s VRPC – with the 32bit RedSq modules – faster – sound works.
Have to softload RO 5.19/5.21 – as you must start with the compressed RO4 rom.
Ref below – perhaps there is a way of making a squeezed RO5 rom.

This was on the web – with respect to the Select rom:-

bash> tail -c +21 Select.rom > Select.Z
bash> uncompress Select.

bash> tail -c +21 Select.rom > Select.Z
bash> uncompress Select.

So that’s saying take the compressed Select.rom file chop off the first 20 bytes and save the result as Select.Z then uncompress all files in the current directory ending in .Z¹ ie. the Select.Z file which is decompressed to Select

¹ Provided the file begins with the correct magic number

Steve R is wrong about the ownership of 4/6 – it’s 3QD that bought the ROL remnants.

Thanks for that – I wasn’t certain who had snapped it up (hence putting the braces and question mark in there).

I suspect the value Steve suggests, and the price Aaron would want (if he was even willing) might differ somewhat.

Indeed. The same goes for Castle. The problem here is that there’s almost certainly going to be a big difference between the real value of these assets – by whatever methodology we select – and the views held by the current owners, who may struggle with the idea of writing-off potentially large investments that were made historically, or they may have wildly unrealistic views of hypothetical potential future earnings.

A hard-nosed view is that RISC OS has close to zero commercial value (present and future) and that the only income these parties can make is by a) selling products that include the IPR (3QD) or b) selling licences to people (Castle). I can’t believe either company is making more than a couple of thousand pounds of profit per annum doing this. In fact, I have a hard time imagining even that much in the way of sales!

Personally, I believe we should Open Source the lot, which should inject some more momentum into the developer community and maximise the chances of RISC OS growing in a way that allows the emergence of an actual, sustainable market around it. RO4/6 isn’t particularly relevant to this, IMO, but I do wonder whether the current Shared Source arrangement is holding things back – it’s no secret that I’ve always had that reservation.

We’ve seen several people shy away from RO just because of the licence terms, if we are to take their postings at face value. (I think they are being unreasonably negative, but their perception is difficult or impossible to overcome.)

I think we’re in a position where RO is just a hobbyist OS. Linux has got the embedded market sewn up, likewise Windows and to some extent Mac OS for the desktop market.

It’s not even a hobby for many people, sadly, although the Raspberry Pi and its recent predecessors seem to have brought a few people in or back.

Of course, to remove the OEM part from the licence (and then get a licence closed to the BSD/MIT one) would be cool. But how much? Castle could also sell the IP to another company, but IMHO, it’s not needed.

We’ve seen several people shy away from RO just because of the licence terms, if we are to take their postings at face value.

That’s a pretty big “If”.

I think we’re in a position where RO is just a hobbyist OS.

Yup. Not only do we lack a “killer app”, we are missing a lot of what people have just come to “expect”. Dropbox? Google Drive? Google Docs? WiFi? Bluetooth object transfer?
This isn’t to say that RISC OS is bad; things are not automatically rubbish if they fall out of the mainstream. This is to say, though, that this is not a competition that we are going to win. So in respects like that, the burning questions is “should we even try” or might it be better to carve out a niche for RISC OS.
I think we have application in the embedded world where Linux is too damn heavyweight. I think RISC OS Pico is an interesting examination in how small a functioning RISC OS can be. What sort of minimal system requirements are necessary? Referring to my PVR (yet again ;-) ), it carries a 16MiB FlashROM and 64MiB RAM (of which 31MiB is available to the OS, the rest is set aside for the DSP code to use for video encode). This system ran Linux with NanoX. When it came to upgrading to a better looking UI with Linux and Qt4, it needed more. A lot more. So the CF slot was claimed and used as OS storage. But, wow, think of what you could do with RISC OS in such a meagre spec. If you compress the OS down to ~3MiB, that would leave around 12-13MiB free in the Flash that could hold a stripped-down version of !Boot, maybe some other stuff. With that much RAM kicking around, there’s plenty of room. A lot could be done…

It’s not even a hobby for many people, sadly,

There’s more money in writing fart apps. What can I say? We’re doomed…

Not only do we lack a “killer app”, we are missing a lot of what people have just come to “expect”. Dropbox? Google Drive? Google Docs? WiFi? Bluetooth object transfer?

http://www.riscosports.co.uk/dropbox.zip

There is even not enough people to port all C, Python, Perl existing code that could permit to have all the needed applications (without really coding). I want FastDosBox2 and an Amstrad CPC emulator :)

Yup. Not only do we lack a “killer app”, we are missing a lot of what people have just come to “expect”. Dropbox? Google Drive? Google Docs? WiFi? Bluetooth object transfer?

It’s tough if you can’t take advantage of developments on other incompatible systems. Things were hard enough in the mid-1990s. When I think about migrating from Linux-based systems to something else, I consider how much work it will take to build a reasonable foundation of libraries and applications that I could use. These days it seems to be a good idea to have a good POSIX compatibility layer if you want to keep up with the outside world.

This isn’t to say that RISC OS is bad; things are not automatically rubbish if they fall out of the mainstream. This is to say, though, that this is not a competition that we are going to win. So in respects like that, the burning questions is “should we even try” or might it be better to carve out a niche for RISC OS.

In a discussion yesterday, I recalled that the 8-bit Acorns became less interesting for me when the 32-bit ones were released, but now I see the attraction in writing software for them. Maybe RISC OS needs to become a full retro platform to find its niche.

One opportunity that I think has been missed is the educational angle. Back in the 1990s, education was a nice niche for RISC OS but it held back the platform in other ways. However, now there is a huge back catalogue of educational software and a fairly new piece of educational hardware to run it on (in principle). Maybe the gap between the demise of Acorn in the educational market and the rise of the Raspberry Pi in education was so long that none of the software producers are still able to take advantage of that.

I think we have application in the embedded world where Linux is too damn heavyweight. I think RISC OS Pico is an interesting examination in how small a functioning RISC OS can be. What sort of minimal system requirements are necessary? Referring to my PVR (yet again ;-) ), it carries a 16MiB FlashROM and 64MiB RAM (of which 31MiB is available to the OS, the rest is set aside for the DSP code to use for video encode). This system ran Linux with NanoX. When it came to upgrading to a better looking UI with Linux and Qt4, it needed more. A lot more. So the CF slot was claimed and used as OS storage. But, wow, think of what you could do with RISC OS in such a meagre spec. If you compress the OS down to ~3MiB, that would leave around 12-13MiB free in the Flash that could hold a stripped-down version of !Boot, maybe some other stuff. With that much RAM kicking around, there’s plenty of room. A lot could be done…

I think it’s probably widely recognised by now that your PVR provider didn’t manage to do the best upgrade possible. I occasionally got to see code that people sent to the Qt support team and it wasn’t always the best quality. Plus, you could strip down the features in the libraries to reduce their size by quite a bit. I sympathise with the developers: it was probably not their job to know the library inside out and to know how to get the most out of the fewest features, but it could be challenging.

One opportunity that I think has been missed is the educational angle.

What education means can cover a lot of ground. That did not matter in the early days of Acorn, because everything was new. But now the educational angle could be from a variety of viewpoints: that of the seller of software, the hard-pressed teacher, a student new to programming, a postgraduate needing to learn about research tools, … etc. These all have different, sometimes incompatible, requirements. Should a scientifically literate person know something about programming? If yes, how much?

Our faculty never found satisfactory answers to these questions. When we tried to involve and get the advice of big businesses all discussion was drowned out by their disagreements over details – they were hopeless.

I can see only one educational role for RISC OS: teaching the fundamentals of computer science, by pushing complexity into the background. I do not count learning how to use certain bits of software – e.g. spreadsheets, Mathematica and so on – as part of that; however useful it may be to learn to use these tools for purposes of demonstration.

A lot of the educational resources for RISC OS were support materials for other subjects. My brief brush with that was related to software for teaching history, or for reinforcing the key points learned in history lessons. This sort of software falls between the extremes of teaching computer science (or just programming) and teaching basic IT skills (which shouldn’t really be necessary these days).

So, much of the value in the software produced was that it supported something that was outside computing. Whether it’s now outdated due to curriculum changes is another matter.

@ David:

It’s tough if you can’t take advantage of developments on other incompatible systems.

True. I think that porting code is quite a bit harder than for other systems exactly because the only thing we have right now (that I’m aware of) is UnixLib.

These days it seems to be a good idea to have a good POSIX compatibility layer if you want to keep up with the outside world.

POSIX, or more than that? Desktop applications (which is, let’s face it, the majority of things) will require something more. Something, perhaps, akin to MinGW and/or Cygwin.
Unfortunately this is where things get difficult. MinGW is not tied up with GPL restrictions, but I’m not sure if the core library itself is open source. Cygwin is open source, but unfortunately it is of the GPL variety. The MinGW website points out the difference thus: Unlike other ports of GCC to Windows, the runtime libraries are not distributed using GNU’s General Public License (GPL), You, therefore, do not have to distribute your source code with your programs unless, of course, you use a GPL library in your programs.
MinGW is not POSIX, it is enough to build working applications using the Windows API instead. Cygwin is POSIX.

Maybe RISC OS needs to become a full retro platform to find its niche.

Just do *Desktop. There’s retro right there. ;-)

none of the software producers are still able to take advantage of that.

Ha. Haha. Nyahahahahahaha!

Sorry. I’m trying to get ahold of some source codes to long-forgotten programs from around the turn of the millennium to bring some more things to RISC OS 5. It is shocking how much code is “on a disk somewhere” or “consider it lost, I’ve movd house four times since then” or even “RISC OS? you’re still using that?¹”. Most of the RISC OS educational catalogue predates this. I would love to see more software, and where I can I’m obnoxiously badgering people, but I don’t hold out much hope for a lot of things. It’s a shame.

your PVR provider didn’t manage to do the best upgrade possible.

I think they stripped out about as much as they could while keeping the stuff that was necessary.

I occasionally got to see code that people sent to the Qt support team and it wasn’t always the best quality.

Enjoy: https://svn2.neurostechnology.com/linux-r3-main-app/trunk
[note: different provider to that used originally and the certificate wasn’t changed, you might need to set a temporary exception if your software moans]

Gavin:

Should a scientifically literate person know something about programming? If yes, how much?

It depends upon the field. Astronomy, for instance, uses a lot of custom software – from hardcore image processing to even knowing what to do with radio telescopes and the data they receive.
People that like making gadgets. You can do a lot with logic gates and transistors. You can do a lot more with a PIC or AVR.

When we tried to involve and get the advice of big businesses all discussion was drowned out by their disagreements over details – they were hopeless.

There’s your mistake right there. Involving big business.
Big business doesn’t give a crap about you, or what you want. They’ll be your best friend and promise you the moon on a stick if it looks like an opportunity to make money. If it looks like an opportunity for some other business, they’ll raise as many objections as possible just to keep their competitor from profiting at what they aren’t going to.

I do not count learning how to use certain bits of software – e.g. spreadsheets, Mathematica and so on – as part of that;

Indeed. I wonder if one of the main barriers to the adoption of Linux in business is that the staff do not know what a word processor is, they know “Word”. They don’t know what a spreadsheet is, they know “Excel”. Hundreds of things may be done, but you give them a different package with a different layout and maybe slightly different terminology, they’ll be lost. Because they don’t know, they just know how to on X.

It is a double-edged sword, mind you. For most of what people were being taught – in my age group, learning WordPerfect (early) or Word (later) was plenty sufficient. There was no point in doing anything more – the boss doesn’t want somebody who is intelligent, the boss wants somebody who can make form letters… So education aims to deliver that. I’ve seen some pretty miserable things in my life (Slough, for instance) but nothing compares to the abject misery that I felt inside when I walked past the secretarial course at college. Thirty odd girls, all sat in front of computers typing out what was being dictated from a tape recorder. It’s like this is it, girls, this is what your job will consist of. Transcribing the mumblings of an idiot far better paid than you. You will meet a guy who doesn’t really love you because he hasn’t ever really figured out if its is you or himself that he wants, and he will buy you appliances for Christmas. I mean, nothing quite says “I love you” like a food processor or a steam iron, right?²

¹ This person, who will remain unnamed, expressed quite some measure of surprise when I pointed them here and said it is more or less open source and it runs on modern ARM boards like the Pi.

² …I’d rather die. If/when I ever get around to doing “the girl thing”, I vow never to give an appliance as a Christmas³, birthday, or anniversary present.

³ Tanabata, Yule, whatever festival is “the yearly one”.

About twenty five years ago a neighbour’s son, whom I knew to be mad keen on computers, came to me for some advice about his education. His school had given him a choice: he could either do an A-level in Computer Science or he could do an A-level in Ancient Greek, but not both. I told him, in no uncertain terms, to do Ancient Greek. I told him that the more useless his knowledge the greater its cachet and the more likely he would become a manager than a serf. All pure Thorstein Veblen (The Theory of the Leisure Class).

And what is he doing now? Running his own software company, of course.

I told him that the more useless his knowledge the greater its cachet and the more likely he would become a manager than a serf.

Senior managers like to have people under them that “think”¹ the same way they do and at the same level.
People who they know to be a genuine threat to their security of bum on comfy seat are plotted against². Evidence of wrongdoing, real³ or imaginary, is gathered to make the case.

¹ OK, “think” is stretching it a bit in some cases.
² You sometimes get an itch between your shoulder blades.
³ I think answering “you’re making me look stupid” with “Oh, I can’t claim the credit for your hard work” probably sits halfway.

POSIX, or more than that?

What we need is the ‘configure’ command (autoconf).
UnixLib is not so bad. Less problems than with the so-called Posix library of BeOS.

“RISC OS? you’re still using that?”

Or “you should use RISC OS 4”. Pfff.

Should a scientifically literate person know something about programming?

France (long time ago) and USA consider that everybody should be able to develop software. The same way we write books and not only read them.

I always promote programming for all. Of course, a lot of people don’t care. But when they have the basic knowledge, you see that they make batch files, macros, and other simple things.

Then I was young, everybody learn coding at school. Of course, some people complain, but they all managed to make working code. Today we learn how to use Internet Explorer.

all managed to make working code.

In the days before microcomputers I think the average output for students at Sussex University who had to code was four lines of FORTRAN per annum.

I always promote programming for all.

Good for you. I see it as learning how to express yourself with precision. It may well be that French students are better taught than English in this regard. I have long argued the prediction that in the long run the best service that computers will have rendered mankind will not be their practical uses but the philosophical consequences of engaging with their stupidity.

POSIX, or more than that? Desktop applications (which is, let’s face it, the majority of things) will require something more. Something, perhaps, akin to MinGW and/or Cygwin.

Probably, but isn’t RISC OS still missing some of the basics, like threading and process handling?

Sorry. I’m trying to get ahold of some source codes to long-forgotten programs from around the turn of the millennium to bring some more things to RISC OS 5. It is shocking how much code is “on a disk somewhere” or “consider it lost, I’ve movd house four times since then” or even “RISC OS? you’re still using that?1”. Most of the RISC OS educational catalogue predates this. I would love to see more software, and where I can I’m obnoxiously badgering people, but I don’t hold out much hope for a lot of things. It’s a shame.

I’m more interested in things from before that time but I’d be interested to know what you’re looking for.

[PVR]

Enjoy: https://svn2.neurostechnology.com/linux-r3-main-app/trunk

Thanks. I can’t see how they configured the libraries from a quick search through that, so there’s probably a script somewhere in another repository. The code there seems reasonable enough for what it has to do – building a useable UI from fairly low level components isn’t easy – and they used the embedded version of Qt, so that saved the cost of running an X server. Maybe they added too many features to that version above what was in the NanoX version.

I think answering “you’re making me look stupid” with “Oh, I can’t claim the credit for your hard work” probably sits halfway.

I’ve already mentioned before. My English teacher asked me, in the middle of class, “Do I look like an idiot to you?”.

Hey, it was in front of everybody. If one is going to fall to the bait, might as well do it good. :-)

The same way we write books and not only read them.

There. Fixed that for you.

Or am I the only person that finds Sartre to be incomprehensible?

Then I was young, everybody learn coding at school. […] Today we learn how to use Internet Explorer.

That’s when Computer Studies (learning how to use and abuse a computer) turned into Information Technology (learning what a computer “does” (but practically nothing about how it does it)).

In the days before microcomputers I think the average output for students at Sussex University who had to code was four lines of FORTRAN per annum.

Well, before microcomputers, getting anything written, compiled, and results back was an epic trek no less than dragging elephants across mountains.

but the philosophical consequences of engaging with their stupidity.

I wonder how long it will take to realise that Facebook was a really bad idea…

The problem is, stupidity is often overlooked when the results are beneficial. Everybody and their uncle knows that Google are starting to get a bit evil nowadays – but we use Google because Bing is crap and Yahoo murdered the rotting carcass of Altavista. We use Google Maps because nothing else comes close, especially street view (I walk around Japan a lot when I have nothing better to do). Apple Maps is kind of cool when it can reconstruct Baltimore in 3D, but this facility only exists for a handful of cities so it is just a toy. Google Docs has torn down pointless barriers between my equipment so the same file can be viewed and edited on pretty much everything I own (RISC OS excepted) without the usual pain of trying to shift files from one device to another. My Tracks is fun/scary to have running when you’re at a boot sale. You feel horrible and want to get a burger? Well, of course you do, you just walked five miles at 1.3mph average and it took you……. And then there is Google Translate. It can still make some rather comic messes out of Japanese, and it still won’t negate in French if “pas” is present but the “ne” is omitted. But it’s the best thing we have by far, and it is the best coverage we have by far. Google image search is also pretty good, so long as you don’t ask it to suggest “similar looking images”, with no recognition of what the image is, the suggestions can be comically wrong. And… well, we’ll just ignore Google Goggles. It used to be good and provide a running real-time recognition of items, but now it takes ages to “scan” a picture and it still can’t tell Mont St. Michel from the publicity photo…
As you can see, there are clear benefits to using the services offered, and we do so in the knowledge that some profile held upon us is being refined every time we search for things, do things, go places… Evil? Perhaps. But it is an evil that benefits. It might not be wise in the long run, but to be honest there are enough other problems in the world to worry about.

Oh, and Amazon, I know you are profiling/tracking me through embedded adverts on other sites. Please, for the love of <Deity$Name> stop advertising stuff I’ve already just bought!

Oh, and Amazon, I know you are profiling/tracking me through embedded adverts on other sites. Please, for the love of <Deity$Name> stop advertising stuff I’ve already just bought!

I get this one a lot: “Thank you for buying $X on Blu-ray. Would you like to buy $X on DVD?”

In one case I bought a two-pack of ink cartridges and it offered a one-pack. Just what I always wanted!

What we need is the ‘configure’ command (autoconf).

GCCSDK?

Yep on UNIX, but no way to use ‘configure’ from RISC OS. And to be honest GCCSDK is very difficult to install/use under Linux. Before I was using ‘configure’ for x86, make a few adaptations to the generated make file, then cross compile (just by changing path to GCC). Now it seems to be much more complex. A guide would be welcome.

Just out of interest, why do people want to use threads on RISC OS?
Threads (with a few exceptions) seem to be a way of fudging a badly designed program (possibly badly designed for a reason) into doing what it shouldn’t be doing in the first place, causing pain and suffering along the way.
(Following comment is not targeted at anyone on this forum….or to do with RISC OS or anyone I currently work with. Not all people appreciate my humour)
I’m not saying all Java programmers are hopeless and have no concept of anything to do with concurrency ;-)

Just out of interest, why do people want to use threads on RISC OS?

I think there are two main reasons:

1. Threads are easy to use and a well understood method of concurrent programming (but not necessarily the most efficient or safest method)
2. Threads are what pretty much everyone else uses – and so if we supported them, porting code from other systems would be much easier

I don’t really like threads, but I can see goo uses of light threads. Light threads could be implemented in RISC OS. We already have the PMT engine of the task windows module.