Post a Comment
The price tag of the Raspberry Pi is 25$ and 35$ (us dollars, not euros)
source : http://www.raspberrypi.org/?page_id=8
Yes, yes, YES. I, too, have always wanted to experiment with RISC OS, but as noted, most of the options involved way too much cash for a hobby machine. From the outside looking in, RISC OS appears to be a very interesting OS - Amiga-like in its well-integrated GUI paradigm.
I would also like students to interact with the computer at a lower level (see command-line interaction in the article). If we only teach them to point and click, they become servants of the machine. If they learn how the machine works, and learn from the beginning to interact at a lower level, the machine becomes their servant.
I've always felt there would be two classes of people in the future - those who tell machines what to do, and those who are told what to do by the machines.
Well, such computers were already used in education. In fact, wasn't this the biggest market for ...RISC OS? The model was probably even more apparent with earlier BBC Micro, or Apple II ...but I'm not certain if they resulted in widespread insights which you envision. Maybe/possibly a bit the contrary (and people with lasting skills chose such path[1])
Anyway, "well-integrated GUI paradigm" of RISC OS might be a bit frustrating, IMHO & possibly[2] - feels way too dependant on drag'n'drop (that's frustrating even now, with optical mouse and reliability of switches mostly worked out - it must have been a nightmare on the mouses from 2 decades ago, especially considering the "public" ones were, typically, permanently clogged)
1. You know, the whole idea of specialisation which gives us our modern world - ultimately, machines are just prostheses of our bodies, also minds ...and most of the fields of knowledge which you touch daily as a layperson (not knowing much about them, just using them) have some machines in the chain crucial to their efficient functioning, I guess. Interwoven into the working of our civilisation.
I even strongly suspect one machine tells you when to wake up, most of the mornings
2. Judging from ROX Desktop (BTW, not requiring any cash), which supposedly does strive to imitate the core concepts of RISC OS GUI.
Ahh, but why did you allow this authority to the machine in the first place?!
Of course, because you (we) suck at it, because that particular machine has much better "timekeeping intelligence"
And consider: if those machines (possibly synchronised with the network of other machines - where Skynet lurks
), one time, were to be mistaken or manipulated... most likely you would still listen to them most of that morning :>
(now add up all other machines on which we depend, usually hardly realising it; that is the whole point of ~AI, mass production and distribution of expertise; and - how large part of "our" economy revolves around... building and maintaining machines, anyway? ;p )
Generally (and trying to drag the topic in a serious manner
), I don't feel it changes much that we give away "control" more or less wilfully - after a while, we forget about the old ways of doing things, we push them aside (and, one could argue, it frees the limited capabilities of our minds to focus on other tasks ...as long as we don't have super-intelligent machines able to outstrip us all in everything, that is
)
Traffic lights are a hilarious example:
1) We certainly are being told what to do by a machine.
2) "But, but, those who programmed the machine know what they are doing" ...well, I'm not quite sure about that. :>
Every single time new traffic lights show up in my neighbourhood, the flow of cars grinds to a halt, and half of the city (the jam "spills over" to other intersections) is paralysed during rush hours (or even outside them, whole day or two!), for a few days
(until "they" figure out, largely by trial and error it seems, which of the presets work decently; traffic & lights planning is largely done by specialised software anyway, so the group "in control" is very small - heck, I suspect large part of their models come from simulations, automated analyses of large sets of ~"traffic recordings", evolutionary algorithms, and such - in a way, quite depending on machines in the search for solutions ...and all this so we can efficiently operate machines which, by _their_ requirements, totally hijacked layouts of our cities and the style of life
)
I think this was always a matter of taste more than anything. Back in its day RISC OS was very much mouse driven, even more so than Mac OS in some ways. In my opinion the advantages (e.g. the best window management and menu system in any GUI) more than made up for issues like that, but I can see how it could frustrate.
Through added features and 3rd party utilities RISC OS became much more keyboard friendly as time went on. Most modern RISC OS applications support a global clipboard rather than relying on drag and drop to move data between applications.
It's the underlying OS that looks primitive today, not the user interface.
Taste is certainly a part of it, but I also think it's a bad design - when the action is interrupted, not only you have to start it from the beginning, it can often bring unpredictable results (when dropping something where you didn't intend to).
From what I saw when mini-training few people over the years, in the use of ~WIMP UI & mouse - it's a bad idea expecting from an average person, right at the start, to routinely do a coordinated, complex (two things, while aiming at some different GUI element), continuous action with a mouse (and again, I shudder to think what the experience would be - even to experienced users - on a typically-always-clogged public ball mouse)
It's good they, as you say, added some alternatives over the years.
Back in those times (after not-quite-RiscOS Arthur), I think Workbench 2.0+ was also nice - and much more widespread, in a way (though I don't think most of Amigas ever ventured outside of the "gaming / toy computer" area)
Well, it may be to you. But guess that most people don't want such a primitive OS that is completely written in ARM assembly (and therefor has nearly no hope of getting any modern OS features)
RPi Foundations intention is to bring modern computing to children for cheap and that is most likely if they get Linux running properly on those machines and at the moment proper X drivers really would help with that.
RiscOS might be great for your short-lived toying fancy, but it sure as hell isn't important in the grand scheme of things.
I'm all for alternative OSes but they better use the full capabilities of the computer.
3D acceleration is needed for what exactly? It's not like you are going to run Cryengine 2 or Unigine on it, anyway. I think it has linux drivers with 3D acceleration, as they were able to run Quake 3 Arena on it.
Indeed, I mean accelerated 2D. Currently, all drawing is done by software on a framebuffer. Moving a window across the display makes my eyes hurt. I'm not arguing that the device is useless without 2D acceleration, rather that it's a much smoother experience with it. And if the device can do it (maybe by using 3D to accelerate 2D if it lacks dedicated 2D hardware) then the software should employ it to that purpose. IMO that would be a much worthier goal than running N other OSes.
When idealized (and towards what present hardware architectures - maybe also R-Pi - converge to, internally), sure.
But, you know, not only 2D / video is something traditionally, for a long time, done by fixed-function, specialised bits of hardware (and libraries, methods, commands oriented for 2D rendering), without any serious possibility to tap those capabilities for 3D.
And not only there are still very clear vestiges of the mentioned separate libraries, methods of accessing 2D vs. 3D.
Even with a HW which doesn't really distinguish internally, and with a compositing WM which directly taps 3D (which Openbox/LXDE doesn't do) - there is still big difference between a stack suitable for "real 3D" (say, the gaming heavyweights mentioned above), and one suitable for desktop (here, OSS drivers are mostly good enough ...even if needlessly requiring more powerful hardware for "good enough" performance; even if, via barely supporting power saving, they waste power - which wouldn't really be a problem for R-Pi, anyway)
Yet when running RISC OS on the same device moving and scrolling windows is super smooth and fast, and that has no acceleration either. Maybe you should be asking the LXDE developers why their code is so sub-optimal!
First you would need to ask the framebuffer or VESA(or whatever driver they use) developers, not the LXDE devs.
Second RiscOS is handtuned assembly and it sure it does a few things faster than rudimentary compatibility code.
But being written in assembly makes it nearly impossible to extend and keep up with the fast changing world around us.
You have to decide, do you want a horse that only goes fast in a straight line or do you want a horse that actually goes around corners?
Wait, seriously, a horse analogy, while cars would be perfectly fine?
(seeing how Clarkson semi-constantly highlights this difference between American and European cars)
PS. I guess it ultimately boils down to how the graphics stack used by LXDE assumes, optimally, exploiting basic capabilities of even "ancient" graphics hardware (I know a functioning driver makes a huge difference even on my Matrox Millenium), fallbacks being there "just in case".
While RISC OS is from another era, assuming lack of such luxuries but, as you say, being able of much less.
Edited 2011-11-01 08:23 UTC
Parts of this could change! http://www.riscosopen.org/forum/forums/5/topics/731
Edited 2011-11-01 12:01 UTC
Amusingly the RasberryPi processor started out as a gpu. They then bolted on a cpu (on the same die?). That is why it has such an unusual boot up sequence and such an underpowered application processor (at least in comparison with most arm contemporaries today).
Edited 2011-10-31 15:03 UTC
Oh my Thom, yes I jolly well do!
This is the most exciting news in ages for me. I started programming on a BBC Model A (that's right A, not B) and have always loved Acorn's OS's. LIke you though I could never justify the price of a dedicated RISC OS machine. Now though...
Raspberries and RISCOS and Linux, oh my!
This is really great. I was a long term RISC OS user back in the 90th. While people switched from the C=64 to the Amiga, I bought an used Archimedes 3000, after reading a special edition of germanies biggest C64 magazine, covering the Acorn machines and RISC OS.
As the machine died I got myself an Acorn RiscPC, including a 486 daughter board. Running DOS from inside WIMP was way cool. I used it for BBS purposes mainly.
Well, after switching to Linux and selling my Acorn Hardware I occassionally thought longingly about RISC OS.
The Raspberry Pi is a nice place to start with RISC OS again, although I'd love to see it running on an ARM based netbook.
This is fantastic news. I've been trying to find a machine to run RISCOS on for ages.
I love the line :
"This will force users to interface with the machine in a programming manner and think of it as something to be commanded at a fundamental level rather than just used."
I knew there was a fundamental difference between computers now and computers as I remembered them but could never put my finger on what it was. I think this is it.
Luckily, RISC OS camp realized (even if it took some time?
) it's a bad idea to jeopardize any remaining chances your niche platform might have, by focusing on niche & expensive hardware ([hint]"Amiga"[/hint]
)
Should be fun, we'll see how it goes (even if "The price point, educational focus, and software support are exceptional" - emphasis mine - is IMHO a bit premature at this point; those aspects might still end up like with OLPC & Sugar...)
Arguably, the RISC OS camp didn't realize it, they just got lucky by relatively cheap and generic ARM hardware becoming popular for some things.
The Amiga camp moved to PPC back when PPC was viable, and there were, IIRC, a couple attempts to move to Macs just as the PPC Mac was dying. Now, it's not a popular platform for this sort of thing, so they have to make their own hardware - just as the Acorn community did 5-10 years ago.
Perhaps, though RISC OS camp does seem to jump relatively quickly on inexpensive almost-retail hardware; they saw their chance in targeting such machine.
The Amiga camp did it way late (excluding AROS, but it needs its time). Only recently MorphOS runs on some old, "surplus" Macs, while Powermacs were available for quite some time. And I think it's more appropriate to discuss those, whole machines, when it comes to past viability; not simply the PPC CPU - which anyway was, for many years, essentially just a coprocessor.
(the whole movement seemed more about few small manufacturers feeding on the leftovers of Amiga following, to keep them alive a bit longer; while perhaps there were more optimal routes - consider how Amithlon, in its days, smoked pretty much every "true Amiga" hardware, for much lesser price, and IIRC it could provide a gradual & transparent to the OS shift towards x86-compiled code; but, the community had strong ideological issues with x86... ~"well, Apple doesn't use them, they use what's best" turned out funny)
Of course, for a long time, there were three kinds of RISC OS machine:
Old RiscPCs with somewhat faster CPU cards (or just the original CPU cards - those are the elderly that keep the commercial RISC OS world afloat)
Newer custom-designed machines, sometimes with their development partially subsidized by part of it being part of an "ARM Linux devkit" that's low-volume, expensive hardware (one of those, with a far slower CPU than the Raspberry Pi, is still available - the better Iyonix was killed off due to RoHS issues)
Windows machines running Virtual RiscPC (and one company is still selling those) - and these get the "but it's not ARM" treatment
They were pretty quick about jumping on the Beagleboard, though, when it became available. Actually, it was the "ZOMG WE NEED A RISC OS LAPTOP" mentality (the only production RISC OS laptop had an ARM3, so it was woefully out of date) that led them to a port to the Pandora, which led them to the Beagleboard as their dev platform.
I use to love RiscOS. My first computer was a Acorn A3000 and I stayed with Acorn until the RiscPC StrongARM.
BBC BASIC was my first programming language and I wrote games for myself, friends and family. It was on the RiscPC I learnt C++ (with gcc). Most advanced thing I wrote on it was a realtime software renderer (all fixed point of course) for a demo/exercise.
BUT even at its birth, really, RiscOS was a toy OS.
"The OS is single-user and employs co-operative multitasking (CMT). While most current desktop OSes use pre-emptive multitasking (PMT) and multithreading, RISC OS remains with a CMT system. Many users have called for the OS to migrate to PMT. The OS also has rudimentary memory protection, and all users have full superuser privileges."
http://en.wikipedia.org/wiki/RISC_OS
Much as I loved RiscOS, I would much prefer kids learnt Linux. It's a true grown up OS and what they learn will be directly useful at many scales. Fundamentally a Unix is simpler than RiscOS. Everything is a file and system calls are to act on those files. (Yes it's not quite that, (bar Plan9) but that's the general idea.)
Python is better than BBC BASIC ever was. Moving to C and C++ is easy on Linux, fancy IDE or not (and I didn't have one on RiscOS, I just had !Zap).
If your also a RiscOS refugee, just use RPCEmu, it can at least run your old 26bit code.
I keep a loose tag on the RiscOS world, mostly in case something better than RPCEmu (good code, but always lacking the love it deserves) comes along to run my old programs, but also because it was once home.
I also am following the Raspberry Pi project as even with the lack of open graphics drivers, and the scary GPU binary blob boot, I may still have uses for a few. It's also fantastic to try and get kids programming. Linux is a fantastic learning platform and very empowering. You know, the old: Teach a man Photoshop and he will be able to edit bitmaps for a day, teach a man Gimp and he will be able to edit bitmaps for a lifetime. ;-)
I really don't want any energy taken away from Linux work on this device for RiscOS. None. If you want RiscOS chances are you really want the GUI, in which case, get stuck in with the Rox desktop.
It's a nice effort but unfortunately it doesn't change the UI of any of the applications you run, which still behave like Windows/Mac/KDE/Gnome apps, so you lose a lot of the benefits of the RISC OS GUI.
Every time I see a File Open/Save dialogue box I just want to cry at the massive waste of time (mine and the programmers) that they are :-(
Agreed. ROX doesn't provide anything close to the full RISC OS experience. Just like adding a dock and theme to Windows and Linux can't turn them into Mac OS X.
Yer, the load/save was the best in any OS. I've felt for some time the desktop should do load/save dialogs and processes communicate with the desktop via pipes/socket. That way you can have Explorer like extensions without injecting code into every process that opens a file dialog. It also means you can do things like proper Rox file dialogs for all apps without them need to even be recompiled. Should be part xdg standard. :-)
I don't see his comment like that, don't really feel your question (as stated) is warranted, Thom - jabjoe was one of the relatively few RISC PC home users, and clearly quite into it; he stayed that way for a while. His insight is valuable / why it needs to be only praise?
(and I don't even agree with him about RISC OS drag'n'drop overuse nostalgia ;p ...especially since the bearable parts of it are mostly available in modern systems)
To me, teaching students with RISCOS is like teaching them Latin -- it's not something they are going to actually use in the real world.
I'd rather have them learning in Linux, which is a mature OS you can command at will, while still getting to use your computer when you want to actually get things done.
It's like schools teaching students LOGO or BASIC. How many programs you actually use in your job were written in LOGO or BASIC? Even Visual Basic is looked down on by true software developers. I say start students early in Python, C and C++.
That's kind of how I feel. But I probably feel Latin is more useful to know than RiscOS. Unix is Latin really. Understand Latin and you understand many languages. Understand Unix and you understand many OSs. Linux is just the Latin all the cool young kids are speaking today. Ok mixed metaphors but you get my point! Python and Linux over BBC BASIC and RiscOS by a thousand to one!
That's mine too, but I just want to be absolutely crystal that this should not take one second away from time spend working on Linux. Any time/energy/money taken away from Linux in favour of RiscOS is harm as far as I'm concerned. I do kind of feel the RiscOS guys should be joining in with Linux to be honest. RiscOS is over.
*sigh*
We're not talking about RISC OS taking over the world, or competing with Windows, Linux, or Mac OS X in any serious way. That's not the point.
However, that doesn't mean it can't be fun. Saying everyone should just use Linux and forget anything else even exists is an elitist attitude that can only lead to stagnation. As far as desktop use goes, Linux is also completely useless compared to Windows and Mac OS X - should everyone just focus on Windows and the Mac and forget about Linux?
We're not talking about RISC OS taking over the world, or competing with Windows, Linux, or Mac OS X in any serious way. That's not the point.
I agree. What's the point in reading a news site about all OS' if you're only interested in one? Or even three? RISC OS is an irrelevency, few sane people would deny it, but why does that mean people should stop developing for it? Pot plants are nice, but irrelevent; should we then all grow cereals instead?
I'm interested in new OSs. OS developments and yes, random OSs like good old RiscOS. But I don't want Raspberry Pi guys spenting one minute on RiscOS because it should be about teaching and empowering kids. As I've already said I feel the RiscOS fans would be better improving Rox than nurse RiscOS any more. Put a RiscOS like GUI on Plan9 design with Linux hardware support and that would be my perfect OS. :-)
I used to be a RISC OS user and programmer, but have since moved on to other things.
Let me be clear - as an OS, technically, RISC OS is a heap of crap. People are indeed not going to use RISC OS in the real world. That is not the point.
This news is a good thing. Not because the Raspberry Pi will run RISC OS, but rather because it means that you'll be able to easily write and run ARM assembler on them. There's no hoops to jump through, no additional software to install, no complicated toolchain to learn - ARM assembler is built into BBC BASIC, which is at the heart of RISC OS.
That is pretty damn useful, and potentially is something that some may eventually have a use for in the real world.
And whilst I was not a fan of Latin lessons at school (I had to take it for three years, and at the time hated it) I do now recognise that there is a big advantage in understanding Latin when it comes to understanding the many other languages it has spawned and influenced.
It's like schools teaching students LOGO or BASIC. How many programs you actually use in your job were written in LOGO or BASIC? Even Visual Basic is looked down on by true software developers. I say start students early in Python, C and C++.
Why do you think schools teach LOGO or BASIC?
It's not about learning a specific language dialect.
It's about learning principles of programming.
If you properly understand the programming concepts you were taught in LOGO and BASIC then those are applicable to Python, C, C++, Ruby, JavaScript, and so on.
The advantage in starting with LOGO and BASIC is that those languages are simple. Much simpler than Python or C.
Python really isn't much more complicated than BASIC, and >much< more powerful. Where I work many Maya artists have been learning Python as their first language and they love it. It also gives them a spring board out of just Maya scripting. BBC BASIC was where I cut my teeth and I have no doubt a younger me would have equally been able to start with Python. Yes it's about the fundamentals but why not start out with a useful language to start with? BBC BASIC was a useful language on RiscOS after all. If it was just a toy language, it wouldn't have kept me interested. Seamed like everything was done in BBC BASIC and ARM code. Python is a modern day equivalent but is cross platform and used on a scale BBC BASIC never was. If you want to do assembler no doubt there is a module for that. There is a Python module for most things after all. Assembler on native Linux is easy enough anyway.
"
@filename: hw.S
.text
.align 2
.global _start
@ ssize_t sys_write(unsigned int fd, const char * buf, size_t count)
@ r7 r0 r1 r2
_start:
adr r1, msg @ Address
mov r0, #1 @ STDOUT
mov r2, #16 @ Length
mov r7, #4 @ sys_write
svc 0x00000000
@ int sys_exit(int status)
@ r7 r0
mov r0, #0 @ Return code
mov r7, #1 @ sys_exit
svc 0x00000000
.align 2
msg:
.asciz "Hello Android!\n\n"
"
Then:
as -o hw.o hw.S
ld -o hw hw.o
./hw
On the user side, its actually quite sophisticated. It was among the first OSes to offer scalable fonts and anti-aliased text as built-in features rather than as third-party add-ons. It was one of the first systems to offer "bundle-based" applications (a la OS X). It was one of the first systems to include a dock. It made use of things like context sensitive popup menus from the start. Indeed, when playing around with RISC OS in emulation, the UI actually holds up quite well even today. On the implementation side, however, things are not so pretty. While it's somewhat cleaner than Mac OS classic, it's nowhere near up to the standard of NT-based Windows or any of the various Unix-likes on the market. Still, given that it's small, fast and efficient, I can't see why RISC OS would be a bad fit for a small computer like the RP. It's not like folks are going to running servers or anything on it...
I have several systems, mainly Alphaservers/Alphastations running Linux and the amount of useful software is really low. Besides of Blender and Firefox there are not that many. Most are either half finished or unrefined not suitable for real work. Amount does not equal quality.
From all the different systems I have collected along the years, must say the IYONIX is really one of the most fun and usable ones I have.
Gimp, Inkscape, LibreOffice, Apt, Hugin, Handbreak, VLC, XBMC, I could keep going. As a RiscOS fan, you must know of Artworks (which kind of carried on from !Draw), which became Xara Xtreme on PC, which Inkscape UI is based on. Inkscape is practically family. I've used Linux for my home desktop for years and I'm happier with it as a desktop then I have been since I left RiscOS. Guake gives me F12 command line goodness (and keeps the desktop alive) and Bash/Unix shell is better than the RiscOS command line shell. Python is better than BBC BASIC ever was. The only thing I still miss of RiscOS is the save/load drag drop.



