RISC OS Open needs your help | Drobe.co.uk archives

RISC OS Open needs your help

Published: 3rd Oct 2006, 22:38:14 | Permalink | Printable

Ask not what RISC OS can do for you, etc

Web monkeys and source code jockeys are being recruited by RISC OS Open to help coordinate the 'shared source' RISC OS 5 project. Recruit is used loosely here as ROS Open say the open source-ish initiative will be run purely on a "voluntary basis". The team is hoping to hear from charitable people who know their makefiles from their wildwildwikis.

Company secretary Steve Revill asked for anyone "interested in helping out with the RISC OS shared source project" to visit the RISC OS Open website.

Specifically, ROS Open is seeking generous website programmers skilled in HTML design, Ruby on Rails, Radiant CMS, CVS, Subversion, Collaboa, type, RForum and Instiki. They're also after suitably competent software engineers who will give up their free time to review the quality of submitted source code and to maintain the operating system's documentation.

To enlist, see the links below.

• ROS Open have updated their FAQ to explain why they decided against using the GNU General Public Licence, and how they drew up the initial list of components they hope to release.

According to the company, "The process of opening up access to the RISC OS sources is a tricky and time-consuming one. The further down into the core of the OS you go, the more bits you find in there that mention or relate to confidential projects from the past and present.

"So we have had to make a choice about what we can release first with the least effort in order to judge the reaction from the community. Also, we wanted to select components which would potentially be of immediate use."

They also hope to get the operating system building with GCCSDK as well as the Norcroft C/C++ package.

Links

Step right up folks, step right up RISC OS Open and Castle announced 'shared source' ROS 5

Previous: RISC OS 5 source code release revealed
Next: How to create a modern desktop theme

Discussion

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

I don't need people; RISC OS needs people.

riscosopen on 4/10/06 12:52AM
[ Reply | Permalink | Report ]

I've put my name down to help in any way I can. Hopefully I'll have a bit more time than in the past 18 months now I've got my PPL

druck on 4/10/06 9:25AM
[ Reply | Permalink | Report ]

Ask not what RISC OS can do for you, but what you can do for RISC OS.

Ginger2 on 4/10/06 10:54AM
[ Reply | Permalink | Report ]

Well, I think RISC OS should run on top of Minux.

Then you'd get a stable, modern, secure kernel with all the features RISC OS's kernel lacks. You'd inherit all sorts of advanced tools from the unix world, including debuggers far better than !DDT, and ports of FireFox would be far easier.

It's along the lines of what Apple did with MacOS, and it's transformed the OS to a leading-edge design.

Or, at the very least, someone should make (most) modules run in user mode.

Piers on 4/10/06 9:23PM
[ Reply | Permalink | Report ]

Sigh. Minix, not Minux.

[link]

Piers on 4/10/06 9:34PM
[ Reply | Permalink | Report ]

I agree, with RISC OS on top of a unix kernel it could be a serious contender. Or to put it more correctly without doing that it is not a serious contender.

DavidPilling on 5/10/06 5:06PM
[ Reply | Permalink | Report ]

Of course, Acorn considered doing this in the early 90s - on top of a BSD kernel (before my time). RISC OS Gold, I think it was called (each RO release had an internal name based on a colour). But the machines didn't have the memory at the time - that's the main disadvantage, really, but memory is very cheap these days.

These was going to be a compatibility window so RO Wimp apps could run - the wimp's message system really enforces co-operative multitasking. Again, this is similar to Win16 emulation on Windows, and MacOS Classic emulation on MacOS X.

Piers on 5/10/06 7:57PM
[ Reply | Permalink | Report ]

Unix does Unix very well.... would Unix do RISC OS well ? I don't think so.

If I want to use Linux I'd fire up Red Hat 7 on my PC at home - why bother with this RISC OS pretense at all then.

Reading things like the above just get me tired beyond belief. Linux has a place, Windows has a place, is *no one* going to speak up on RISC OS's behalf - or is it just to be treated as another means to coax/cojole or otherwise push the few remaining users onto another platform.

We have feet guys - we're not obliged to shoot them

AMS on 5/10/06 8:03PM
[ Reply | Permalink | Report ]

Why all this talk about porting RO to another platform. One of the main problem,in my opinion, is the lack of software development in general. Not many new applications have been announced/updated since Acorn called time on RISC OS. This I believe is the biggest problem with the platform, not whether or not it runs on the latest version of the most 'in' geeky Unix or similar kernal.

sa110 on 5/10/06 9:11PM
[ Reply | Permalink | Report ]

The thing about a Linux kernel is that it makes it easier to build the superstructure that allows for easy porting of Open Source applications. Linux has open source applications, Mac OSX has open source applications helped because of its kernel. Windows has Open source applications because people decided to put the effort into porting to the defacto standard. What Open source applications do you see for Risc OS? An old version of Firefox that still incomplete. A version of Gimp that needs X Windows Client again an Alpha release. Your not going to see lots of open source applications unless you make life easier. So you are going to miss out on the Open Offices, The Gimps, Scribus Inkscapes XaraExtreme, Firefox, Thunderbird Ekiga, plus other apps like Skype

Jwoody on 5/10/06 9:28PM
[ Reply | Permalink | Report ]

The thing about RISC OS at the moment is the lack of true multitasking (much debated over the years but this aspect of Windows is much nicer), and its fragility - poke the wrong place and the whole thing falls over - not what you want in an embedded application. However if you make RISC OS less fragile, in even the simplest way, you will break most of the existing desktop applications and there are not developers around to fix them.

DavidPilling on 5/10/06 9:42PM
[ Reply | Permalink | Report ]

Exactly. What RISC OS needs more than anything else is compatibility with other platforms. The Rest of the computing world is working hard to become more compatible with all other platforms, except for RISC OS. Even Microsoft has realised this trend and has heavily invested in its "Linux and open source compatibility lab". The key do doing it right is to retain the RISC OS GUI and introduce a compatibility layer for legacy apps.

JGZimmerle on 5/10/06 9:50PM
[ Reply | Permalink | Report ]

Jwoody: not true. What we have is precisely most of this functionality to make ported programs possible. That's thanks to the GCCSDK development and the UPP, and all the people who've contributed to that. What we _lack_ is developers and motivation for the developers who are still here. The RISC OS version of Firefox is not old, indeed it's more or less the current version.

David: I'm not sure I fully agree. There are a couple of simple things that can be done to RISC OS to improve its stability which STD pursued with Adjust, such as zero page read protection. That breaks a couple of things, but those are easily fixed.

mrchocky on 5/10/06 11:21PM
[ Reply | Permalink | Report ]

mrchocky> STD's current ROM image does not even protect zero page from USR mode writes, never mind reads. This is something that I sincerely hope will be changed soon.

adrianl on 5/10/06 11:46PM
[ Reply | Permalink | Report ]

I'm not sure what made RISC OS great anymore, a few killer apps from ComputerConcepts (now surpassed by Scribus, Xara etc.) but mostly the WIMP I guess, certainly not the stability of the OS or ease of programming desktop apps.....

I dunno, I can't see much of a future for RISC OS on ARM hardware, or maybe even RISC OS as we know it (kernel, WindowManager etc). Maybe porting to x86 may save it - VRPCSE has been popular I guess, but native would be nicer than emulation.

Perhaps just doing a ROX kind of thing and porting the Desktop to Linux may be a way to go, imagine the RISC OS desktop running on a stable OS - multitasking, memory protection, being able to use whatever peripherals you like, a huge library of applications available for free, current hardware..... ROX never seemed to catch on though.

Thing is, if you make a Linux desktop or Windows emulator, you're just going to convert people to Linux/Windows in the long term, and they fire up ROX/VRPCSE less and less.

simo on 6/10/06 12:15AM
[ Reply | Permalink | Report ]

Obviously a contentious issue...

Modules have absolutely no memory protection; Applications have very little. Even something mundane such as wimp applications can easily crash the system - overwrite a menu pointer or window icon string, and you're lucky if you only destroy another application's window, but usually the entire OS falls down in a heap as one heap is shared by critical kernel tasks and application window workspace.

This is because the kernel was never designed, and was just cobbled together to meet a deadline. Acorn always planned to replace it (ARX, RISC OS Gold, TAOS (Intent), Galileo).

The good parts of RISC OS have nothing to do with the kernel. The font manager, GUI, filer, image filesystems etc could all be ported to a modern kernel, keeping the feel of RISC OS, but at the same time adding pre-emption, speed, stability, and creating a much easier platform to develop under - especially when writing drivers.

And comparing a microkernel such as Minix to RedHat is misguided. Minix itself is probably smaller than the existing RISC OS kernel. I'm certainly not suggesting the GUI runs on X11.

Piers on 6/10/06 12:45AM
[ Reply | Permalink | Report ]

One thing I don't understand, is why people can not understand that using another system's kernel for RISC OS does not mean replacing RISC OS with that other OS. The kernel is only a very small part of the OS and one the user never directly interacts with. All user interaction takes place in things like command-interpreters (the CLI) or GUIs (the WIMP).

And while we are talking about utilising a different kernel in RISC OS: Everyone seems to agree that one of the top reasons for taking this step would be to get better hardware support. So why would we even consider using anything else than the Linux kernel? Linux has by a long way the best hardware support of any operating system ever developed.

JGZimmerle on 6/10/06 3:06AM
[ Reply | Permalink | Report ]

in reply to GZimmerle: "Everyone seems to agree that one of the top reasons for taking this step would be to get better hardware support. So why would we even consider using anything else than the Linux kernel?"

simply the 'invasive poison' of the linux licence (GPL). Always a matter of opinion, but a licence that requires you to give your competitors your 'goodies' under almost all circumstances does not help innovative development.

jb on 6/10/06 8:26AM
[ Reply | Permalink | Report ]

jb: yes, that certainly explains the success of RISC OS compared with the dismal niche that the Linux kernel operates in.

Jaffa on 6/10/06 9:14AM
[ Reply | Permalink | Report ]

The GPL is neither "invasive" nor "viral", and you would do well to avoid such language. I agree it's not very satisfactory wrt RISC OS modules, but I still urge you to consider a compatible license.

mrchocky on 6/10/06 9:59AM
[ Reply | Permalink | Report ]

mrchocky "What we lack is developers and motivation for the developers who are still here"

If true why not publish a guide to producing an executable

1) Where to get a Risc OS version of CVS client 2) Where to get and how to install GCCSDK and UPP 3) Where to get and how to install CHOX11 ( Or what ever its called ) 4) How to get source tree from CVS for GIMP say 5) What to change in Makefile to setup Libraries and external variables etc 6) type make 7) What to look for in compiler errors 8) What to look for in linker errors 9) What needs including in !run file

I am sure that if people could at least get to produce an executable they could run, even if it crashes. It might start them on the road to porting. I fully agree there is lack of experienced developers around. But without some guide people face a very large learning curve just to get started and I think that puts off a lot of less experienced people

Jwoody on 6/10/06 10:39AM
[ Reply | Permalink | Report ]

Jwoody: I think I would have to agree with this. I'm a 25 year professional, commercial applications developer, but not in C, so starting something in C would be incredibly difficult without a lot of starters. Certainly something like thismight help me get started.

Dave

DS1 on 6/10/06 1:07PM
[ Reply | Permalink | Report ]

If Castle setup their C/C++ library with scripts to build RISC OS, it would be another incentive to consider purchasing

markee174 on 6/10/06 1:23PM
[ Reply | Permalink | Report ]

IIRC the GPL would not "infect" RISC OS modules, as they are "separate works".

Here are the relevant paragraphs:

---------- snip ---------- These requirements apply to the modified work as a whole. If identifiable sections of that work are not derived from the Program, and can be reasonably considered independent and separate works in themselves, then this License, and its terms, do not apply to those sections when you distribute them as separate works. But when you distribute the same sections as part of a whole which is a work based on the Program, the distribution of the whole must be on the terms of this License, whose permissions for other licensees extend to the entire whole, and thus to each and every part regardless of who wrote it.

Thus, it is not the intent of this section to claim rights or contest your rights to work written entirely by you; rather, the intent is to exercise the right to control the distribution of derivative or collective works based on the Program.

In addition, mere aggregation of another work not based on the Program with the Program (or with a work based on the Program) on a volume of a storage or distribution medium does not bring the other work under the scope of this License. ---------- snap ----------

This basically means, that if an identifiable section of source code also works without the GPLed parts, then it is considered a "separate work" and the GPL does not automatically apply to it. Since RISC OS' modules are identifiable sections of code wich would work without other GPLed modules and without a GPLed kernel, for example when compiled for the RiscPC, they would not automatically fall under the GPL.

JGZimmerle on 6/10/06 1:59PM
[ Reply | Permalink | Report ]

Super digest follows...

DavidPilling: "I agree, with RISC OS on top of a unix kernel it could be a serious contender. Or to put it more correctly without doing that it is not a serious contender."

Wise words indeed.

AMS: "would Unix do RISC OS well ? I don't think so."

Mac OS X does classic Mac nicely enough, so there's a notable precedent that you're ignoring.

sa110: "Why all this talk about porting RO to another platform. One of the main problem,in my opinion, is the lack of software development in general."

The two things aren't entirely independent. Having the machine lock solid because some kernel module (erm, relocatable module) dips into the OS workspace or some hardware register really undermines any attempt at a reasonable development environment.

simo: "I'm not sure what made RISC OS great anymore, a few killer apps from ComputerConcepts (now surpassed by Scribus, Xara etc.)"

I'm not vouching for Scribus, and Draw still does some things more elegantly than Inkscape, although that's mainly due to the inability of developers to maintain usability whilst adding features, but there isn't any part of the RISC OS experience I can think of that isn't surpassed by something like KDE plus GNU/Linux. Back in 1994, RISC OS could have taught the UNIX desktop scene a few tricks; now the influence is definitively reversed.

JGZimmerle: "This basically means, that if an identifiable section of source code also works without the GPLed parts, then it is considered a "separate work" and the GPL does not automatically apply to it."

This is why large parts of KDE, including derived works based on KDE code, can be distributed under the LGPL, although a complete, functional KDE distribution would most likely be licensed under the GPL due to the underlying library licences.

guestx on 6/10/06 3:18PM
[ Reply | Permalink | Report ]

guestx

"and Draw still does some things more elegantly than Inkscape"

For an equivalent of Artworks or better see XaraExtreme but then they both eminate originally from Computer Concepts as was. Certainly more elegant than Draw and reads Draw files to boot.

Jwoody on 6/10/06 3:24PM
[ Reply | Permalink | Report ]

JGZimmerle: My reading of the first paragraph contracts the meaning you are assigning to it. Separately distributed GPL softloading modules do not affect the status of other modules or the kernel, but a distributing a ROM image "a whole work" comprising of some GPL modules and/or a GPL kernel would come under the clause; "the distribution of the whole must be on the terms of this License, whose permissions for other licensees extend to the entire whole, and thus to each and every part regardless of who wrote it."

GPL is the wrong licence for RISC OS. Its aim is to foster collaborative development to compete against the established commercial interests of first Unix then Windows, and to prevent that work from being bought up and suppressed, or used without permission in commercial products. Its good for that purpose, but is not the "god license" and by no means suitable for every purpose, no matter how willing people are to evangelise it.

While we want the release of RISC OS source to encourage a open source developer community feeding back enhancements for the benefit of everyone, that is not the sole aim and will not ensure the future of RISC OS. To ensure that both hardware manufacturers and software developers stick with the platform, there still needs to be financial input from commercial ventures using RISC OS in vertical markets. For RISC OS to be considered 3rd parties must be able to make their own enhancements, adding value to the product without having to make that code available to their competitors, as the GPL would require.

Lets not get lost in the license issues - over in Linux most effort is now creating noise over GPL2 vs GPL3, and we can't afford to waste time like that. If you are interested in the future of RISC OS then what’s on the table is a reasonable compromise between commercial and open source interests, and will allow us to start working and taking the OS forward.

druck on 6/10/06 3:40PM
[ Reply | Permalink | Report ]

druck: "over in Linux most effort is now creating noise over GPL2". I watch Linux development very closely, and I'm certain that's not the case. We're much better at wasting time in RISC OS land, and wheel reinvention.

"GPL is the wrong licence for RISC OS." Let's not be so quick to make such sweeping statements. GPL might well be an excellent choice for parts RISC OS. But note that I was very careful in my wording and said "GPL-compatible".

mrchocky on 6/10/06 3:54PM
[ Reply | Permalink | Report ]

jwoody: [link]

mrchocky on 6/10/06 3:55PM
[ Reply | Permalink | Report ]

mrchocky: "I watch Linux development very closely, and I'm certain that's not the case."

Indeed. Relicensing Linux might now be conceivable (thanks to the auditing which followed the SCO charade), but isn't a particularly likely event. Comments from prominent kernel developers have mostly been posturing and/or misrepresentation of the actual GPL 3 content, but I doubt that they're diverting anyone's time away from technical matters.

mrchocky: "GPL might well be an excellent choice for parts RISC OS. But note that I was very careful in my wording and said "GPL-compatible"."

Moreover, a dual-licensing arrangement where the GPL is offered alongside the "proprietary licence for paranoid companies" would present a superior solution to that proposed. Nobody is really suggesting that the GPL would placate those paranoid companies, so stating the situation as the GPL alone vs. the RISC OS "Open" combo (as druck does above) really adds no insight to the discussion. All I stated in the comments to the original article was that for the stated aims of RISC OS "Open", the GPL would be superior to their "in due course" licence for the reasons given by mrchocky in that same comment thread (and alluded to above).

guestx on 6/10/06 4:16PM
[ Reply | Permalink | Report ]

IN reply to anyone getting their heads round running RISC OS on a linux kernel.

Back in my PhD days I used to run an automated framegrab from a microscope running MacOS8. For a wee bit of stability for the weeklong experiment I uised Mac On Linux.

see [link]

Now before anyone says emulation, back then it was more simple. I turned on the machien and it booted straight into MacOS8. So I assume RISC OS would be the same. If ported in this way.

Cheers Bob; back to Slackware/VectorLinux.

nijinsky on 6/10/06 4:18PM
[ Reply | Permalink | Report ]

mrchuck

"in reply to jwoody: [Link: www.riscos.info]"

Thats the sort of thing but NEEDS to be more !!!!!! than just GCCSDK

It needs to have

1) Where to get a cvs client 2) Where to get ChoX11 3) How to request GIMP source tree with CVS 4) Changes required to Makefiles for ChoX11 libraries

i.e. Take somebody through the whol process

Jwoody on 06/10/06 4:32PM
[ Reply | Permalink | Report ]

druck:

"My reading of the first paragraph contracts the meaning you are assigning to it. Separately distributed GPL softloading modules do not affect the status of other modules or the kernel, but a distributing a ROM image 'a whole work' comprising of some GPL modules and/or a GPL kernel would come under the clause"

No, the distribution in binary form is not what is meant by that paragraph, as is clearified in the last pragraph I quoted from the GPL. So my point still stands, that we could use the Linux kernel as a new kernel for RISC OS, _without_ the risk of having to put legacy RISC OS modules or applications under the GPL as well. To make that absolutely certain, we only have to ensure, that the modules and applications (in source form) will still work with the legacy kernels as well.

I am certainly not evangelising the GPL, it is just another licence. However it would be useful to bring Linux' hardware support to RISC OS. With the RISC OS sources becoming available, it will become a lot easier to make native builds of RISC OS for different hardware architectures, once a cross-assembler and an interface-layer between a Linux kernel and RISC OS modules has been created. An ARM emulator could be integrated into non-ARM RISC OS versions that provides compatibility for legacy applications wich are not developed anymore.

JGZimmerle on 6/10/06 4:42PM
[ Reply | Permalink | Report ]

How about trying what's there, before demanding more information or making assumptions. Unless you can give it a try, there doesn't seem to be much point. And your '!' key seems to be stuck. However:

(a) Included in your unix environment (b) the autobuilder will build this for you. (c) I could tell you this, but it wouldn't do you very much good. More useful would be the Gimp in the autobuilder. I can easily do this, but that doesn't mean it works (it doesn't, work is required on freetype) (d) none.

Due diligence and some patience is required on your part. Asserting what we "must" do will not get you anywhere fast. Start with the wget example then go from there.

mrchocky on 06/10/06 4:50PM
[ Reply | Permalink | Report ]

Hmm. I like how someone mentions Unix and everyone assumes Linux.

I suggested Minix for a number of reasons. Firstly, the licence is compatible, but secondly, the Minix kernel is technically far superior to Linux - it is a proper microkernel. The Linux kernel is historically a large monolithic kernel, with some drivers considered part of it. Drivers run in supervisor mode, so offer little protection. Things are changing, but the underlying design is not neat.

The downside is that Minix has very few drivers, and doesn't even support the ARM yet...

Piers on 06/10/06 7:03PM
[ Reply | Permalink | Report ]

mrchocky

So are you going to supply autobuilder scripts for all Open Source Applications? Or are you expecting users to build their own?

If I was a linux user and wanted to work on GIMP say I would go and get the cvs source as per the instructions on the GIMP web site and then run make having made sure I had all the prereq libraries.

If for example a user wanted to try build XaraXtreme for Risc OS do they make a request via you to have an autobuilder script created.

Jwoody on 06/10/06 7:13PM
[ Reply | Permalink | Report ]

Let's actually try some examples first. When you have, we'll see you on the GCCSDK list, and we'll be happy to answer your questions.

mrchocky on 06/10/06 7:42PM
[ Reply | Permalink | Report ]

Piers:

I did not automatically assume Linux. I already knew quite a bit about Minix (through my education) and Solaris x86 (I used to be a Solaris admin), and since you mentioned Minix I read up on current developments. I agree, that Minix is a neat design, but I don't believe it would be a good idea to use it as a foundation for a mainstream OS, simply because the developer base is too small and it supports too little hardware. By merging RISC OS and Minix, we would create yet another technologically superior OS wich won't stand a chance in the real world.

JGZimmerle on 06/10/06 7:47PM
[ Reply | Permalink | Report ]

mrchocky "Let's actually try some examples first"

Well I inherited some machines at work with cygwin on Windows installed on them at work. Such a bad experience I am never going near cygwin again. Thats sort of leaves Linux or Mac OSX neither of which I have. If I win the lottery ( which I rarely enter ) perhaps I will buy a new machine and put Linux on it.

Still another vote for Risc OS based on Linux at least one could build open source applications for Risc OS without needing a Linux, Mac OSX or windows + cygwin

Jwoody on 06/10/06 8:03PM
[ Reply | Permalink | Report ]

Jwoody:

Well, if you have a computer that runs Windows (wich you would need to run Cygwin), then you also have a computer wich could run Linux. You can simply install Linux on your PC in addition to Windows and select wich one to start at boot-time. Linux is free (in both meanings of the word), so why not give it a try. I recommend Kubuntu, wich you can either download and burn onto a CD or order a free installation CD to be sent by post from www.kubuntu.com

JGZimmerle on 06/10/06 8:48PM
[ Reply | Permalink | Report ]

I've tried a couple of Linux distros on a variety of PCs and found them slower (to boot up) and less reliable (never got dial-up or broadband) networking working reliably than the versions of MS Windows they were supposed to replace. I've always wanted to like Linux, but always found it, somehow, disappointing.

jms on 06/10/06 9:06PM
[ Reply | Permalink | Report ]

Julian

"Well, if you have a computer that runs Windows (wich you would need to run Cygwin), then you also have a computer which could run Linux"

No chance not enough room left on hard disk. Machine has only room for one. I guess I could add a second via USB but don't have enough money for that. Have to live on retirement pension. Hand round the hat

Jwoody on 06/10/06 9:21PM
[ Reply | Permalink | Report ]

If hard drive space was truly the only barrier to you having a setup such that you could do this type of RISC OS development, I'm sure someone would donate a drive. 10-20GB or so would be fine.

mrchocky on 06/10/06 11:25PM
[ Reply | Permalink | Report ]

In reply to mrchocky and Jwoody:

For someone being used to GCCSDK it is probably dead easy but for someone wanting to try it it is quite a learning curve. I did go to www.riscos.info trying to find out how to get and use GCCSDK.

So below "Compilers" I clicked on the GCCSDK link - after discovering that on riscos.info links are not as usual underlined but just coloured text (whatever the reason to not adhering to the common link display - probably due to using a Wiki). There I then found the "Using GCCSDK" which is a short intro which is surely enough for someone knowing Linux ... to just get started. But what "<path to gccsdk-dir>" is, might have been mentioned and the AutoBuilder.html page mentioned here is unfortunately not a link. But what to do after AutoBuilder is done is not stated (probably just copy the executable to RISC OS).

As for ChoX11 I followed the link from riscos.info main page and did find a bit info on it like that its importance should not be underestimated but then apart from downloads there is not much to find in order to learn how, what for and why to make use of it. Probably not much to worry about anyhow but perhaps a few more words here might help.

I think that some step-by-step tutorial for "newbies", that is one not assuming all sorts of knowledge to start with, showing how to create a simple CLI app like wget and a simple GUI one might help attrating more GCCSDK users with the tuturials starting off with nothing and thus with getting and installing GCCSDK etc. way down to having the final app on RISC OS. Could be that such a tutorial is there but then that should be easier to find on riscos.info.

Or simply put: Quite a bit of information is there but partially not easily found, or with quite some parts left for the potential user to discover, or simply omitted. That fits the "Due diligence and some patience is required on your part." statement of mrchocky well, doesn't it ...?

hzn on 07/10/06 07:49AM
[ Reply | Permalink | Report ]

I think the best way to introduce RISC OS user to a Linux-based RISC OS development environment, would be a bootable Linux CD like Knoppix or Kubuntu with installed GCCSDK and a graphical IDE.

JGZimmerle on 07/10/06 09:33AM
[ Reply | Permalink | Report ]

In reply to Jwoody regarding hard-drive space.

I would organise my files into used-regularly and used-infrequently. Zip the infrequent files. Windows these days runs files from archived formats.

Download Ubuntu-6 or Simply-Mephis-6 (Ubuntu based) for an easy partition and install. Ubuntu is best on existing windows systems but Simply Partiutioner works better.

If you have an older system you would eb better to use Vector Linux. Based on Slackware (the UNIX of Linux) so is VERY stable and it is faster than Ubuntu or Simply. Incidentally it has xfce with ROX (I use KDE, yeah I know bu tit is for work on my laptop and presentations so others find it more familiar).

To not be able to acheive this your hard drive would bneed to be less than 20Gb and since 40Gb has been the standard for years then there should not be a problem. If downloading is a problem, go and buy a linux format mag at £5. This should have at least one distro. Cheers Bob

nijinsky on 07/10/06 10:10AM
[ Reply | Permalink | Report ]

In reply to JGZimmerle:

"I think the best way to introduce RISC OS user to a Linux-based RISC OS development environment, would be a bootable Linux CD like Knoppix or Kubuntu with installed GCCSDK and a graphical IDE."

That is a good idea indeed.

What might be a good starter would be to supply a script file or a package for the odd common distro which installs the full GCCSDK setup with a few sample apps for use by executing that script, or installing that package with the package management tool of the distro. This might be a bit easier to supply opposed to compiling the full CD or ISO image for self-burning. This script should make sure that the odd prerequisite is in place or if it is a package have the needed things as prereq.

hzn on 07/10/06 1:09PM
[ Reply | Permalink | Report ]

hzn: "due diligence". My due diligence towards RISC OS has been paid many times over with documentation, tools and many other things. You'll understand if I don't go of my way to feel bad that not everything I've done is immediately obvious to everyone. Equally, for anyone wanting to get involved with what we have done, the GCCSDK developers are understandably more likely to give help to someone who is both enthusiastic and likely to help themselves, rather than those who like to invent problems. None of that means the things suggested shouldn't be improved, but you need to understand where we're coming from, and how little time we have to spare.

As for a script, I think you'll find that precisely those small sequence of commands on the GCCSDK page (which I just yesterday rearranged to make clearer and as a starting point for others to expand) is almost all of what you say. It installs and builds everything and many of the checks for appropriate tools are done automatically. The fact that it's so short and easy is evidence of this due diligence. Perhaps it's even too easy, meaning there's little to learn - after all, learning is the whole point of others getting involved.

Remember, it's a wiki. Anyone can add things as they see fit, and there's more documentation _inside_ the GCCSDK once you check it out. Being so quick to make assumptions isn't going to help anyone. Asking questions is a lot better.

mrchocky on 08/10/06 01:47AM
[ Reply | Permalink | Report ]

In reply to JWoody:

Would it help at all if a machine were available on the net that you could log into and use GCCSDK?

(It would have to be command line only, but all of the tools could be pre-installed and ready to use.)

gdshaw on 08/10/06 08:52AM
[ Reply | Permalink | Report ]

hzn: Thanks for reporting what you didn't understand in the GCCSDK tutorial - that's exactly the sort of feedback we need. I've expandeed the 'Using GCCSDK' tutorial a bit more. If you try it again please let the GCCSDK developers know if there's anything you still don't understand. (Obviously more depth is still required)

caliston2 on 08/10/06 10:40AM
[ Reply | Permalink | Report ]

In reply to JWoody:

BeatrIX Linux under 200mb's

Surely you have enough room for 200mb?

MikeCarter on 08/10/06 4:21PM
[ Reply | Permalink | Report ]

Try Damn Small Linux or Puppy Linux - both can be as small as around 50MB, meaning they can fully load in RAM. I believe DSL can even run under Windows.

There's really no excuse not being able to run Linux these days.

hEgelia on 08/10/06 4:56PM
[ Reply | Permalink | Report ]

None of these will be practical solutions for GCCSDK. You'll need a "grown up" Linux distribution, with full versions of tools, although the install itself can be relatively minimal. Ideally, that would be Debian or Ubuntu - the latter being particularly easy to install and use. Fedora is easy to install but otherwise an exercise in pain. As for space, to tackle large ported RISC OS programs, I think you'll need at least 1GB free after installing the system, just so you don't run into problems later.

mrchocky on 08/10/06 5:04PM
[ Reply | Permalink | Report ]

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

+++ Message board +++

Read messages or start a new thread

Search the archives

Today's featured article

Online banking with RISC OS
We thought this would be of high interest

28 comments, latest by devine on 25/9/05 8:30AM. Published: 1 Sep 2005

Random article

New site risc-os.de to spread word of ROS
Someone tell CTL and ROL it's this easy

1 comment, latest by Gulli on 4/4/06 12:32AM. Published: 28 Mar 2006

Create a new account
Forgot your password?

Useful links

News and media:
Iconbar • MyRISCOS • ArcSite • RISCOScode • ANS • C.S.A.Announce • Archive • Qercus • RiscWorld • Drag'n'Drop • GAG-News

Top developers:
RISCOS Ltd • RISC OS Open • MW Software • R-Comp • Advantage Six • VirtualAcorn

Dealers:
CJE Micros • APDL • Castle • a4 • X-Ample • Liquid Silicon • Webmonster

Usergroups:
WROCC • RONE • NKACC • IRUG • SASAUG • ROUGOL • RONWUG • MUG • WAUG • GAG • RISCOS.be

Useful:
RISCOS.org.uk • RISCOS.org • RISCOS.info • Filebase • Chris Why's Acorn/RISC OS collection • NetSurf

Non-RISC OS:
The Register • The Inquirer • Apple Insider • BBC News • Sky News • Google News • xkcd • diodesign

© 1999-2009 The Drobe Team. Some rights reserved, click here for more information
Powered by MiniDrobeCMS, based on J4U | Statistics
"Leaking tit bits creates rumours, confusion and often mangles facts"
Page generated in 0.2322 seconds.