RISC OS wish list

Fundamental Improvements

• 3D hardware support (2013-08-25) - 103 - agree - disagree - follow up
  • + yep, like 3-dimensional hard-disc and 3-dimensional floppy-disks or even 3 dimensional key-boards
  • + I agree, but only some sort of driver-interface, so that 3rd party developers can write the drivers. Writing drivers for such things should not be RISCOS Ltd's task.
  • + Most 3D hardware uses OpenGL - A module interface would be needed to allow easy acces to the OpenGL calls from software.
  • + I'd only put it in a module if it's possible to export the interface directly to programs to avoid the overhead of decoding the SWI call etc. The only reason for putting it in a module and not DLL is to support BASIC, which I think is pointless for high speed graphics applications.
  • + OpenGL is not always used for high speed it can be used to ease the burden of high detail rendering
  • + No no! Please let me access 3D hardware from BASIC! Please!
  • + How about putting OpenGL in a module and providing an SWI to read the address of low level handlers? Then both BASIC and ASM/C/C++ people would be able to use the hardware.
  • + + module for use of strongArm to assist rendering(for those with more than one SA)
  • + OpenGL support would be nice. Heh, imagine playing Unreal Tournament on Risc OS :)
• multi-threading (2013-08-25) - 65 - agree - disagree - follow up
  • + The future as reported by some is in mutiple processors, multy threading is the best way to implement mutiple processors on a OS. My idea is that there could be a set anmount of threads (ie one theread per processor) that would allow the standard co-op multytasking on each thread. Provideing the benifits of a co-op system with that of a multy processor system.
  • + We really, really need this. It will make Internet apps which multi-task nicely so much easier to write.
  • + I added some support for this in the Wimp when I was at Pace - Wimp_StartTask, 0 starts a new Wimp task in the current memory slot, in which you can then start a taskwindow in. Unfortunately the code never got merged into CVS although it's around...
• Hardware independence (2013-08-25) - 64 - agree - disagree - follow up
  • + And when it happens, chamge the name from "RISC OS" to something else. RISC OS on a CISC chip?
  • + presumably 'hardware independence' = runs on anything with an ARM in it (but no VIDC or IOMD), not runs on anything at all?
  • - Goddarnit. I'm not letting those grubby intel home-baked chippies at my OS!
  • + I agree that hardware independance is important and I am glad that it has already started with the viewfinder and the new machines
  • - The only sane thing to port RISC OS onto would be a SPARC based system. But would take years... (ARM@ukGinger.net)
• pre-emptive multi-tasking (2013-08-25) - 63 - agree - disagree - follow up
  • - Cooperative M/T has forced developers to write more stable, effiecient code, Windows PreEmpts (most of the time) and look how lumbering the apps are for it
  • - And have you seen the mess Windoze makes when it pre-empts a CD writer making an autoplay CD?
  • + Pre-emptive multitasking does not make code bad. That's the fault of poor developers, and a market thats driven by 'the need to upgrade' constantly
  • + Pre-empting is not bad. What would be nice is a way of starting a pre-emptive thread under RISC OS...but keeping most as co-op
  • + Am I the only person who likes to get on with other things whilst ChangeFSI lumbers along doing its stuff rather than waiting for it to finish?
  • + Only to put a maximum to a processes timeslice so with correctly written programs the multitasking is stil cooperative
  • + Pre-emptive multitasking in both textmode (like Unix,BSD,Linux,Solaris) and WIMP-mode. Switchable CLI-consoles like Linux
  • + A mixture is better though - see Active Objects in EPOC for a fantastically balanced solution
  • - What happens when Windows forces programs to multitask when they're not ready? (They crash with alarming regularity.) Who better to know when a program's ready to multitask if not the program itself?
  • + I reckon that RISC OS should perform pre-emptive multitasking transparently, logging all the resources that an app is using and paging it in and out at the appropriate moments.
  • + There are OSes that multitask better than RISC OS and are more stable, just look at BeOS. Don't judge features like preemptive multitasking by the poor implementation in Windows.
  • - What we need to do is write co-op code, that polls wimp when it's processing - WRITE BETTER, MORE EFFICIENT CODE!!! DON'T get lazy! -&uquest;
  • + Preemtive Mutltitasking or Multithreading is very elegant. (Microsoft screwed it up anyway...) I think it would be a great improvement to RISC OS. For time critical tasks like CD Recording or MP3 playing or whatever ther could be a Realtime Thread
  • + You can run ChangeFSI in a task window - it was in a Risc User Jan/Feb 1995. It was on the disc with it... Maybe on the nutshell CD? - jwd104@york.ac.uk
  • + Co-operative multitasking can be thought of as a special case of preemtive multitasking in most ways - so there's no need to have two distinct mechanisms as such. It shouldn't be too hard to add this to RISC OS (and they can do threading at the same time).
  • + Pre-emptive multitasking does not stop anyone from having some co-operative mt too. Look at AmigaOS Forbid()/Permit()-calls. (And Disable()/Enable())
  • + This is very important. Especially when writing networking code. Having to write totally tortuously non-intuitive code to do the simplest things makes the whole OS feel outdated and stifles further development. We've got to have some kind of priority on this.
  • + This is very important. Especially when writing networking code. Having to write totally tortuously non-intuitive code to do the simplest things makes the whole OS feel outdated and stifles further development. We've got to have some kind of priority on this.
  • - Better to fine tune the exisiting code, pulling out the odd bug and adding marginal improvements
• Memory protection (2013-05-03) - 61 - agree - disagree - follow up
  • + I remember reading a letter in acorn user some years ago about the Acorn A3000 only crashing 3 in 7 seven years. So page memory protection has the potential to ensure system does not crash.
• very fast startup (some seconds) (2013-05-08) - 38 - agree - disagree - follow up
  • + I find it extreme important to be able to switch the machine on and off very fast, making less noise and powersavings too...
  • + I think there was a program which saved the memory to disc and swaped it all back in for instant bootup (would take time to read 16Mb+ from a HD though)
  • + Yep, and support for sleep mode and instant awake (and LAN Wake, Serial Wake, etc)
  • + It is called Freeze and it can be downloaded from http://www.multimania.com/gusuki
  • - ... and if you want your machine to work then you don't use Freeze ! I STRONGLY recommend against it.
  • + You may *not* know how to use your computer or read (manuals). I like Freeze and I use it *without* any problem.
  • + Freeze (my bady) is a powerful utility, it helps if you learn how to use its power. The manual says all you need to know. - GUS/Arm's Tech
  • + I with the pro's on this one!! So long as you ensure no apps conflict with it, then it's fast and seems more reliable than a normal start up. BTW, speaking of freeze, would it be possible to have it load an obey file after unfreezeing? So that any conflicting apps could be loaded in the tradional way?
  • + You can use the files: !Boot.!Freeze.~Default.!AfterR & !RunTasks
  • + How about having a boot from a FlashROM? Have a FlashROMFS (Like ROMFS on the BBC, only R/W) and have different 'disks' for different parts of the boot sequence! It would certainly be quicker, I think!
  • + So very important. Writing networking apps is tortuously difficult at the moment, and developers will just decide not to bother unless we do something about it.
• Java 2 support (2013-09-14) - 37 - agree - disagree - follow up
  • + Preferably JDK1.3, which is in beta at the moment. With this you could run JPython, which would give great command language. Java (JRE) is needed for web browsers, use OJI to connect it up. It also allows RISC OS to be used for lots of e-commerce like stuff and as a thin client.
  • - Why does nobody write improvements to Java VMs in Java (or something of similar concept) so that updates can be universal?
  • - What tool wrote that last comment? If you wrote a VM in Java, it'd still need another VM to run on.
  • + Not True! The Squeak Smalltalk VM is written in Smalltalk ... which is then converted by the system into nicely portable C. C compilers are coded in C. Java VM in Java is not inconceivable.
  • + Lets get native support for this. Why not build a VM into the OS and have double clicking on class files launch an instance of the VM? Or at least distribute one with the OS and have support for using any VM ...
• POSIX (2013-05-10) - 31 - agree - disagree - follow up
  • + Oh, and someone might want to add POSIX... [A set of IEEE standards designed to provide application portability between Unix variants -- FOLDOC]
  • + This would fix a few of the problems in riscos allowing OS supoport for threads,shared memory, PMT and fork(); , so it would open a whole raft of applications to the os
  • + I've wanted to fork() under RISC-OS for ages. I started putting something together that would almost be a fork() [spoon() perhaps], but it got bogged down in difficulties with windows, as the spawned task would either have to declare a whole new set of windows to the Wimp, or somehow refer to it's parent's windows. Either way, nasty.
• Multi-user support (2013-05-17) - 28 - agree - disagree - follow up
  • + Separate sets of application configuration for each individual user. Provide a file protection scheme so people can have access to the machine without risk of them trashing the HD.
  • - The thing I like about RiscOS is how it lets you into the system. I like the HD lock on the RPC and I use LockDir in school. People who are likely to 'trash' a RiscOS hard drive will either not be able to get past the protection already available or will be in confined quarters already.
  • + Well at least some facility for requiring login and security if RiscOS is to support server type activity
  • + A Multyuser system would be fantastic under RiscOS I would recomend a configuration option to allow oneuser/many users but this option would have to be secure One option would be to use a encryped image fileing system so that users get their own !boot.choices and work directorys
  • + RISC OS will never be a proper multi-user OS. If you want profiles/file protection/shared !Boot it should be possible to just use a network setup.
• multi-processing (2013-05-08) - 23 - agree - disagree - follow up
  • - I'd rather have Pentium support first.
  • - No, transmeta support would be much more useful.
  • - Pentium as we know it will finish at the end of this year. Intel have strong ARM links.
  • + Multy processors is the future for all computeing needs It is strange that current computers run all their bilions of instructions through one little black box
• Colour Management System (2013-05-08) - 22 - agree - disagree - follow up
  • + A colour management system to allow standardised calibration of image input and output devices.
  • + Well you will need this to implement POSIX in any meaningfull way.
  • + Incorporate Cerilica Vantage's capacity for dealing with coloured paper into !Printers, and simulated display of ink colours made available to other desktop apps. (Perhaps a chargeable upgrade).
  • + and we need 48 bit colour capability.
  • + I would like to see the pallete returning to our desktops perhaps as a seperate application as I loved being able to fiddle about with the colours of the mouse, borders and windows
  • - Why have 48-bit colour support when the human eye is not even cabpable of regocnising the entire 24-bit palette???
  • + 48bit (or 96bit) allows more shades, therefore it allows more accurate colour reproduction, expecially when using gamma settings etc.
• Internet Access (2013-04-26) - 17 - agree - disagree - follow up
  • + As long as Risc OS cannot equalize the Windows way of Internet Access, we are doomed. So please, give us The Full Internet.
  • + Hear hear! Exactly my thoughts!
• Security (2013-05-10) - 14 - agree - disagree - follow up
  • + Decent hardware lock / password, multiuser capable ie control access for "guests" - also allows potential to configure for different users
  • + I would even like to see an optional 'console login', similar to portables', before any form of booting takes place.
• Automated Virtual Memory Management (2013-09-14) - 14 - agree - disagree - follow up
  • + This should also have a section in the Task window so you can see what is using it
  • + Perhaps the Green Bar could be split indicating how much is virtual and how much is real
  • - Use of virtual memory is unnecessary. All our apps are teeny weeny, and memory costs peanuts these days. V Memory is slow and unwieldy and should be, at most, OPTIONAL.
  • + Graphics application in particular needs a standard from of virtual memory. (A the moment each one uses a different method.)
  • + Memory should be infinite. VM is the cheapest solution. It is obvious that many programs are faster on RISC OS, because there is no swapping possible.Linux is a good example for an OS with VM that is fast and only uses the disk to save data or when no physical memory is available. Good applications are written effective and optimize memory usage, but many features like demons and preemtive multitasking go hand in hand with VM.
  • - What is the point? Only very few apps require this sort of thing and we already have !Virtual. When it kicks in on NT/Mac/Linux they slow down to a snail's pace.
  • - I meant Virtualise or whatever it it called, not !Virtual.
• Multitasking (2013-05-07) - 11 - agree - disagree - follow up
  • + add Wimp2 functionalities to obtain a real multitasking OS (not just task cooperative)
  • - Why care how it's done? If from a users point of view is doing multiple tasks at the same time it is real multitasking in my view.
  • + It may not matter from a users point of view how multitasking is acomplished but for a developer it makes a great deal of difference. Writing software for an operating system with proper multitasking is not only simpler but allows for better performance of the hardware.
  • + If multitasking is under the control of the OS it would also allow for realtime applications which require a specific QOS from the system. This is impossible to guarantee in a cooperative MT system.
  • + Multitasking should be under control of the R Operating System
  • - When systems force programs to multitask when they're not ready, they crash. For numerous examples, press the power button on your buddie's PC. Who better to know when a program's ready to relinquish the system, than the program itself?
  • + Exactly. A well written program should always be aware it could be pre-empted, otherwise it's broken.
• limes (2013-05-17) - 10 - agree - disagree - follow up
  • + RISC OS should support more fully a range of citrus.
  • - I like Lemons.
  • - Surely oranges are the only fruit?
  • - I like Apples, too.
  • + mandarins!
  • - Do we really want to be turning to citrus fruit when there are just so many good, wholesome vegetables out there?
  • - Are radishes vegetables?
  • - There's something about fruit having seeds and vegetables not, so I guess radishes are vegetables.
  • - Can I be a pineapple?
  • + Vote Pineapple for President!
  • - No, Tomatoes!!!!!
  • - Are you taking the pith?
  • + Who ? Me ?
  • + RISCOS does need increased Limes awareness...
  • + Ah, yes, Jam - that is the way forward. Mumble mumble.
  • + Marmalade! (Lime, Lemmon, Orange...)
  • - Limes? Never! Yams!!!
  • + I once wrote a game that was going to turn into a lemmings clone for the BBC called Limons. (That's because, well, my drawing skills weren't exactly very, well, skilled, and they all ended up looking like yellow/green pieces of fruit.)
  • + Pomegranites - They're modular!!!
  • - What the hell are "limes"? :-
  • - I don't think fruit is necessary, just look at the mess Windows makes of fruit, therefore all fruit should be banned.
  • - Ever noticed how Oranges are full of little nodules :p
• BBC Basic (2013-07-22) - 8 - agree - disagree - follow up
  • + BBC Basic could with some more modern ideas, like classes to handle the wimp
  • + If anybody wants BBC Basic with classes, then just supply me with a modern enough machine and set of uptodate manuals. ARM@ukGinger.net
  • - Who wants BBC Basic with Objects and structures ?
• Transparent built in virus detection (2013-09-14) - 7 - agree - disagree - follow up
  • + Add background virus detection on every file so infections are already detected. coo, I've already done that :)
  • + Smashing idea - Pineapple Software might get a bit peeved though - you'd be undercutting one of their main lines of business :-)
  • + I can always supply anti-virus code to anybody who wants it :)
• Ability to send Apps to sleep (2013-05-09) - 5 - agree - disagree - follow up
  • + ATM programmers can provide a mask so that an app only responds to certain events. Could RISC OS allow an app on the icon bar to be sent to sleep by the user so the app is only sent one wakeup event?
  • + There is a app like this which suspends to disc a single application it was designed for low memory systems (Can anybody remember the name)
  • + get RAMplify by werewolf
• Improve Risc OS Command Language (2013-05-07) - 2 - agree - disagree - follow up
  • + Even poor old DOS has loop control structures. I realise that for backwards compatibility we cannot leap to some super new script language. However, it would be nice to have local variables and local functions as well as system variables. I am not asking for Perl, but Risc OS looks a bit thin compared with what is available elsewhere.
  • - It's called BASIC. An extension to BASIC to make reading OS variables easier is all that is needed.
  • - A Perl port for RISC OS already exists.
  • + It needs something like a Bourne compatible shell
  • - A Bash port exists too.
  • + What about Python? It's a nice, easy to read language which even beginners can program effectively. It's free. It's been ported by Lawrence Tratt. It can be used in an object oriented fashion. It is available on nearly all platforms (don't know about the BBC B though 8-)
  • - Perl does all this and more, the manipulation of system variables etc. in PERL is incredibly easy.
  • - BASIC! Its a nice, easy to read language which even programmers can program effectively. Its free. It is available on all Risc OS platforms. And we all know it.
  • - We've had it for 20 odd years, man! BASIC!!!
  • + Some fo the features of bsah would be very usefull os_clis constucts are a bit poor :-(
  • + Whatever is used, it needs to be available as standard - eg in a New Boot structure. Otherwise an application with a perl (or bash or python or whatever) !Run file is a bit stuffed
  • + Python is lovely - all I had to do to fall in love with the language was see some code fragments, and I thought wow!
  • + Is there a way to get bash scripts runnig on an arc?
  • + try looking at hensa - there's one there I think.
  • - Of course there is an Object Perl out there as well
  • - The only problem I've encountered with using Basic as a scripting language is that running an Obey file sets Obey$Dir, so that you can reference files relative to it. Runing a Basic file doesn't. The olution, of course, is that you have an Obey file that call a Basic file. This multiplies the number of files, though. More elegant would be for *Basic to set Basic$Dir to the path of the file.... Hold on. It does (sort of) anyway. OS_GetEnv returns the command path. Memory falling apart. I've been using this method for 14 years...
  • - there is a tool which provides a var BASIC$DIR like OBEY$DIR. should be available at http;//go.to/icebird
  • + Totally replace the underlying command line with Un*x. Mutli-user, shells, scripts, everything-is-a-file.
  • + Some level of per-task environment is essential. Prefix$Dir (provided by DDEUtils), while useful, shouldn't be necessary...
  • + I would do it for free is someone would let me. It would'nt be that difficult to extend it smoothly.
  • - It's good enough already!
• Crash Protection (2013-05-08) - 2 - agree - disagree - follow up
  • + A little idea I had. What about a crash protect module, working similar to the Watchdog (alt-break), but the app should ideally register for crash protection, issuing various informations to the module. If it crashes the module simply performs a (unix-like) core dump of it's memory, reloads the app and reinstates the memory etc..
  • - like while (1) { do_app; } ?
  • - So... if it crashes you get a core dump and it restarts... and does the same thing ? That sounds like a Bad Thing to me. The alternative of copying memory sporadically doesn't work either because there are external object references that won't be the same. In general, if you want to do a 'retrying' OS then you do it from scratch.
  • + Not exactly, the idea was (say) a Word Processor would pass the area of memory storing the current document (s), after the crash the WP would restart with the document storage memory re-instated (I assume most time when a wp crashes asfter I type the letter 'b', it's not the typing of that letter which has crashed it.
  • + How about a module which filters illegal commands - integrity checks, etc - and infinite loops. The app could be quit, reloaded and the memory dump reinstated? Possibly? (Please, please PLEASE... NO BLUE SCREEN!!!!!!)
  • - Arent we forgetting that RISC OS doesn't crash every 5 seconds like the silly windoze ones?
  • - Nothing like that until there's a *tough* memory protection system implemented. That's because the app could crash corrupting the machine's memory. No kidding: I've got nothing less RiscOS-friendly than !Edit do that, simply typing the wrong replace string (would't even keyboard-reset).
  • + Internal Error: Abort on Data Transfer!
• redefine operating system (2013-07-22) - -1 - agree - disagree - follow up
  • + operating system should be an interface that responds to us as humans. voice recognition, gesture recognition should be added to the input profile. then the configuration of the machine should also be decidable on startup. configurations should include phone, wordprocessor, typewriter, radio,television, dvd player/ recorder,organiser,games console, internet intranet port. on startup we should hear ' good morning . what would like me to be?' we should be able to reply or point to (guesture) the configuration. 'hi, risco. phone please.' multi tasking would obviously help and is essential, buta later development
  • + Riscos has a unique place in british hearts.It is a trojan horse in the world of embedded processors etc. Let's do it, let us change the world before they know it.
  • + Damn! Linux beat us to it :(
• Abstracted workspace API (2013-05-08) - -1 - agree - disagree - follow up
  • + A means of reading the address of operating system workspace. OS_ReadAddressOfZeroPage.
  • - Surely OS workspace is OS workspace - for its eyes only?
  • - Decent programers shouldn't even think of reading things in this manner. To be honest the only thing I can think of is Hardware access.
• A per-task CSD (2013-09-14) - -2 - agree - disagree - follow up
  • + A Currently Selected Directory for each task, rather than a single global one. 'The Desktop' should be considered as a task in this definition, so that pressing f12 takes you to the same CSD as last time you pressed f12. It would need the tasks to somehow register the need for it's own CSD to avoid breaking apps that allow you to change the desktop's CSD from the desktop (eg where [Task/Pin]Filter lets you adjust-drag the switcher icon to a filer window to set the CSD.
  • - DDEUtils already provides this functionality. I wish for people to spell "its" correctly.
  • - A utility called CWDSupport is an answer to your needs, it goes beyond what DDEUtils does (registering of private use for some System variables) and can be downloaded from http://bgilon.free.fr/RISCOS/CWDSupport/main.htm Its use has been described in a past Archive magazine issue .
• Techwriter (2013-05-07) - -3 - agree - disagree - follow up
  • + needs to be able to do simple vector drawing within program
• Decemnt working Internet-software, as capable as Winstuff (2013-05-07) - -3 - agree - disagree - follow up
  • + I love RISC OS, but what I really, really miss, is reliable software to use on the Internet. Something like KaZaA for instance. And a good browser with all available Java-handling, so that I finally will be able to use my Telebanking account.
• Run programs in their own "region" (2013-05-07) - -4 - agree - disagree - follow up
  • + Each program runs in a "region" and every region is controlled by RISC OS. Should a program crash, the "region" is terminated by RISC OS and nothing bad happens. If a program crashes in Windows 95/98/ME, it can freeze the whole computer (leaving you no option but to hit the reset button!).
  • + Just to add to what I said... A "region" could be a place in memory/virtual memory for every program that RISC OS has full control over. If a program tries to write to another "region" of memory/virtual memory, RISC OS will block this attempt and shut down the program altogether (with debugging information of course - and the option to report it!). There are many bad things a program could do which warrants closing it down, including directly accessing the hardware, infinite loops, illegal instructions and so on.
• State of the ART audio support (2013-05-10) - -5 - agree - disagree - follow up
  • + support for audio standards like we expect in home audio
  • - You mean like CD Audio quality? RISC OS could do this a long, long time ago.
  • + But the hardware is still not up to scratch. PC users have Dolby Digital 7.1 support, can play DVD, SACD, SD.... and our MIDI support is still pretty average.
  • + We're getting there, with the Cino PCI sound card now released with AC3 5.1 surround support etc., but we're still a while a way yet.
• Hijack PC Motherboards (2013-05-09) - -9 - agree - disagree - follow up
  • + Would it be possible to design an ARM cpu card to fit standard PC slot 1 motherboards, also replacing the BIOS with whatever minimum required to drive the hardware, and boot the rest from the hard drive? Might even allow dual processor boards to have Pentium card as well.
  • + Yes and we could run RISC OS on a banna too
  • + eew! how about pears? hmm.. 2 RISC OS'!
  • - Yeah, let's hijack the PC boxes! Put in ARM, ooooh then we can paint a BIG green Acorn over the Windoze logo! ...errr... no, let's leave the PC junk to PC's. I personally think that Microsoft + Intel + PC design = junk. let's not do ARM + RISC OS + PC design... it's not broke... so don't fix it!
  • - Linux isn't perfect - I've seen it completely munge over its own hard disc structure reducing the machine to a reck. It never has worked properly since.
  • + You can't just plug such a thing into any old ATX motherboard (BIOS), but you COULD do a deal with a single manufacturer and duplicate the BIOS drivers in the RISC OS HAL - this would be a cheap way of getting an extremely good motherboard
  • - Doesn't Virtual RPC do that?
• Proper file closing Star Command. (2013-05-10) - -10 - agree - disagree - follow up
  • + *Close and *Shut are near-useless, because they risk crashing programs that have had their open files closed. What is needed is a command to list all open files with file handles, and extensions to *Close to allow individual files /handles to be closed.e.g. *Close | This would make life much easier when debugging apps.
  • + Oops, my syntax disagrees with HTML... OK then, "*Close 45" (handle syntax) or "*Close "$.dump" (filename syntax).
  • - Don't write broken code :-) Nobody should ever have to call those commands.
  • + These commands are sometimes necessary, though... ideally, the OS would associate file handles with tasks.
  • + It is ridiculous to say "Don't write broken code". Any programmer knows that anything that can go worng, will go wrong. In particular, BASIC programs in an endless loop need to be killed using ALT-BREAK.
  • + Phrase it as "allow apps to tell the task manager what files/handles it is using so that the task manager can clean up if the app dies".
• Give up on the kernel (2013-05-08) - -11 - agree - disagree - follow up
  • + Take a linux kernel/Mach and put RISC OS on top. You get good filesystem, good hardware support, multitasking, memory protection, processor independence all for free.
  • - [Giggle]
  • - You're sort of missing the point of all this, aren't you?
  • + Well possibly but if the suggestions on this page were all to be implemented I know which approach would take less time
  • - Than use an Intel-PC with Linux...
  • + I would, but the GUIs are not as good as RISC OS.
  • + Talking about the kernel, the Linux kernel is open source so whats to stop Riscos ripping some code from linux (USB, VirtualMemory...)
  • - How about things like 'copyright' ?
  • - Copyright? No problems with copyright just as long as the source of the code is acknowleged! I love open source!
  • - Enough!!! I've sat back on my behind for long enough!! NO MORE OF THIS LINUX STUFF! OK! Sure it's great for a server OS. But as a desktop OS, for real everyday use - please!! It slugs like treacle, software doesn't just have to be OS dependant, but chip dependant too! RiscOS is the way forward, it's greatest single-user. single-processor GUI (maybe OS too) that I've ever used.
  • + The RISC OS underlying OS really stinks! The only thing that makes RISC OS worth using is the GUI. I think that having the RISC OS GUI on a powerful OS and fast hardware would be brilliant.
  • - Yes, OK but if you were to put a RISC OS GUI on linux, you still can't run any of the software written for the proper RISC OS you silly person
  • + This should not preclude developers from porting Linux developement tools or end user products for the benefit of the RISC OS user community. Linux has much to offer in this domain.
• Different colour Text for aggree comments (2013-05-08) - -11 - agree - disagree - follow up
  • + I cant read the green text very well. Can someone make it blue ?
  • + What we need is contrast between the white backgroud and the green so a darker background or a darker green but not both
  • + Maybe a new finish? Mahogany?
  • - I just wanted a red line here, because there wasn#t any yet.... :-)
• Bug removal (2013-05-08) - -14 - agree - disagree - follow up
• Decent Startup Banner and switcher icon (2013-08-25) - -23 - agree - disagree - follow up
  • + Preferably using that bryce rendering !!
  • + Possibly in an easier to customise way
  • - I personally dislike startup baners, and would rather have some way to get rid of the default one.
  • - I personally like the current startup banners in RISC OS
• 64 bit ready for the new StrongARM! (2013-08-25) - -156 - agree - disagree - follow up
  • - It's a 64 bit BUS, fool!
  • + It would be a good idea to give guidelines on how to make 64bit compatable code as the PC is moving towards 64bit it is only a matter of time until the StronARM moves up there to join them
  • - And what processor exactly is going to run this 64bit code that someone is meant to produce coding guidelines for ?
  • + So someone make the 64-bit processor first!
  • - Who said that we must support the StrongARM family by Intel? The new ARM10 with a FPU and VectorFP is much better!
  • - Don't forget about ARMULET
  • + We need to get up to 32bit for the newer processors. (Does anyone feel like prompting arm to release information on how a 64bit ARM would work we could emulate it for fun)
  • - Whatever happened to Armulet? I read one article about it donkeys years ago, and then nothing since.
  • - Anyone know what a donkey's year is?
  • + A good point about 64 bit processor. In the next 5 yesrs I expect memory and storage technology to take quantum leap in capacity. This means a wide databus is needed (i.e. 64 bit).
  • + However, a 64bit bus could be implemented by loading two 32bit instructions at the same time ala ARM Thumb
  • + Nah, we should do an Acorn. Since everyone else is going 64bit, ARM should go 128bit!
  • + nah, then make it easier to use and do a 104bit version, we can upgrade to 128bit later
• 32 bit mode (2013-08-25) - -175 - agree - disagree - follow up
  • + 26 bit code still needs to be supported
  • - And 8bit 6502 obviously
  • - Do people understand that 26 bit applications can't run on a 32 bit OS ?
  • - Do users understand anything?
  • - Not really, it seems.
  • + 64 bit mode
  • + Mu-ha-ha... I spotted that was a joke :-)
  • + The level of user comprehension on Acorn's is far far above that of PCs - try getting the average PC user to explain the difference between 26 bits and 32 bits!
  • + + is there not a way of recoginising 26bit code and then altering it to 36bit just before execution
  • - Yes; it's called 'Magic'. (ie, no, there isn't)
  • + What about a real 32 Bit RiscOS with an emulation for the 26 Bit Mode? Apple did a quite similar thing when they introduced the PowerPC (68000-Emulation).
  • + i agree. given that the development of software is slow under risc-os, allowing 26-bit stuff to be used under a 32-bit os is *essential*, if a 32-bit os were to have any sort of future.
  • + Sorry, but I don't understand. If Window9x can run 16-bit legacy apps, why is it not possible to have 32-bit RiscOS running 26-bit apps? And if I remember rightly Computer Concepts released something called (IIRC) Win32 to run 32 bit apps on a 16-bit OS, would this not be possible on RiscOS?
  • + If RISC OS is to have a future, then we need 32bit code, stop living in the past
  • + The new Strong Arm from Intel will be 32 bit only I believe. We really need to educate coders to not use 26 bit instructions as soon as possible. Even in the latest software you still see people ORRing into the PC to set the condition codes for instance. A little bit of thought now will save loads of hassle later.
  • + 128 bit mode
  • + Not too hard to write NEW arm code 26/32 bit compatible - only difficult to update old code. Can recompile C,C++ etc. BASIC will still work.
• add a new item

Return to wishlist, or acornusers.org.