RISC OS features in plain english | Drobe.co.uk archives

RISC OS features in plain english

By Chris Williams. Published: 14th Jan 2006, 07:08:47 | Permalink | Printable

A brave stab at bringing all the feature lists together into one easy to understand summary

Select 4
The details of what is likely to be in Select 4 was published in October by RISC OS Ltd. It involves the 32bit kernel and operating system produced for the Samsung ARM9 powered AdvantageSix A9 range. Select 4 is hoped to be ready in time for Q2 2006, two years after the release of Select 3.

The features list was mostly lifed from the programmers' logs and is therefore rather on the technical side. Below is an attempt at explaining what the more interesting features mean with a summary in plain english at the end of each section. The text in quotes is from the leaflet handed out by ROL at the 2005 South East Show.

Kernel and other low level work
"All the source is now updated to be 32bit safe. Support for 32bit only processor modes is included. Assembler based components are now built using ObjAsm and a modern makefile."

The kernel is at the heart of the operating system and its job is to act as an intermediary between the application software and drivers that talk to the computer hardware. Whenever a serious error occurs or a device needs controlling, it's typically the kernel that takes over and delegates the task of handling the issue. It has been made compatible with modern 32bit ARM processor cores, opening up the opportunity for it to run, for example, on ARM9 and XScale platforms. The new makefile is part of RISCOS Ltd.'s move towards making their operating system less of a headache to update and improve.

"Kernel has lots of abstraction of its graphics API. It now has no pointer support, abstracted to OSPointer module. Video hardware is now driven by an external module. The only significant change that has yet to be completed is the banking of screen modes. Video hardware abstraction now provided to remove dependency on VIDC and IOMD architectures. VideoHWVIDC provides the entire hardware driver for video. VideoSW provides the software implementation for certain graphical operations."

Previously, the kernel has been tied to the Acorn-era chipset. Now this dependency has been gradually removed, leaving a kernel that can be more easily ported to other ARM based platforms. This includes the graphics hardware driver and mouse pointer support, both of which are in separate modules that can be replaced as desired depending on the requirements of the underlying hardware. As an example, RiscPC class computers will be running the VideoHWVIDC module, whereas A9home users will be using VideoHWSMI - same kernel, different driver modules. A separate module can be provided to implement operations in software that cannot be performed in hardware. For example, the A9home's Silicon Motion SM501 video chipset can accelerate in hardware the copying of rectangular blocks of the screen, whereas the Acorn VIDC in the RiscPC cannot. A separate module could perform this operation in software for VIDC users.

The overall result is that RISC OS becomes even more modular and producing ports for other platforms becomes easier. Replacing and improving individual components and functionality also becomes easier.

According to those close to its development, this is also one of the reasons why 32bit RISC OS 4 cannot use the Castle RISC OS 5 kernel: the way in which the new 32bit kernel and the individual component modules that used to be in the RISC OS 4 kernel communicate won't work with the RISC OS 5 kernel, it's claimed.

"It also now correctly supports LD/ST 'T' and SWP in abort trap mode."

RISC OS provides a degree of memory protection so that if a typical application tries to access a protected or invalid area of memory, it is stopped by what is known as an abort - reported, for instance, to the user as the familiar 'Abort on data transfer' message. Sometimes the operating system may want to be notified when a program accesses a particular area of memory. To do this, the OS marks the area of memory as being out of bounds and waits for the program to use it. When the application does so, the processor will detect the operation, pause the running program and inform the kernel via an abort. In this case, the kernel passes control to a special handler that will silence the abort so that no error message is generated, perform some background task related to the watched area of memory, execute the program's instruction that triggered the abort and then return seamlessly to the application, allowing it to continue as normal. This is used with lazy task switching and accesses to special cached screen areas, where the operating system must intervene if a program attempts to access particular areas of memory.

A similar system exists called Abort Trapping, which works in exactly the same way except it emulates the trigger instruction rather than executing it directly, and device driver developers can use this feature to their advantage: a piece of hardware can be 'mapped' into memory, which means it can be treated as an area of memory where settings can be stored, status information retrieved and commands to be performed written to. Instead of plugging in hardware and mapping it in, during development the device driver programmer can instead opt to just mark the are

Previous: Remote desktop apps compared
Next: News in brief

Discussion

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

Chris - just had a good read through all three sub-articles. It must have taken you ages! Extremely well-researched and easy-to-read, and gives an excellent overview. The RISC OS 4.40 article also does a great job at breaking down some of the newer features into things that non-techies will understand (to an extent - they're still very technical subjects, but certainly easier to read than it was!) Many thanks to all involved on this for a great job. Will make a good reference piece for some time.

md0u80c9 on 14/1/06 9:53AM
[ Reply | Permalink | Report ]

Magic!

After asking for something like this in the comments on the 2006 predictions - & it appears.

Great work guys.

Only thing ROL need to do now is to include options for the folder colours - that green is yak! ;-)

ROHC on 14/1/06 11:07AM
[ Reply | Permalink | Report ]

Brilliant!

I've hoped something like this would appear for a while now, but this is better than I could have expected! Excellent work and this should serve the interests to far beyond just the platform community.

hEgelia on 14/1/06 11:36AM
[ Reply | Permalink | Report ]

And yet it has to be asked - why didn't the people who produced the OS do the same?

This applies to many things - not just RISC OS. Some of the documentation available for other platforms sucks harder than a whore on a saturday night!

Nodoid on 14/1/06 12:09PM
[ Reply | Permalink | Report ]

Nodoid:

I do agree; this kind of thing should be up on the RISC OS Ltd website in particular. Castle do seem to have been slightly better at explaining what RISC OS is all about on their website.

A really enlightening article, and one that's going to be useful to refer back to in future, when people ask about RISC OS. Thanks drobe

flypig on 14/1/06 6:29PM
[ Reply | Permalink | Report ]

Hmmm, been mod'd down. Fair enough.

Put it this way. If, when OS 4 was released this sort of information was around and in such an accessible way, I very much doubt that those who maintained their "I'm sticking with OS < 4" stance (especially those on OS 3.1) would have stayed that way.

I somehow doubt that collection of people realise the long term damage they did by adopting that stance and more over, how little ROS Ltd did to help themselves by not writing as clearly and simply as drobe has here (or even writing much at all).

Nodoid on 14/1/06 7:44PM
[ Reply | Permalink | Report ]

Ok agreed that the various Risc OS players could have done 'their bit' better. But have you ever thought that in their eyes they had? Communication that reads well is very difficult. How many times have we complained that the help files in software is rubbish? Think who writes it.

I take my hat off to anyone who manages to take, what in essense, is a technical subject and make it readable/understandable by 'the man in the street'.

Well done Drobe!

Mart on 15/1/06 12:52AM
[ Reply | Permalink | Report ]

Mart : I agree with what you're saying, but given what we've seen since the inception of OS 4, can you *honestly* say that they have?

Look at it this way - OS 4.0x offered some BIG changes over 3.7 and the difference between OS 4 and 3.1 was so large as to be unbelieved. However, due to a lack of talking non-technical, the uptake of OS 4 was not as big as hoped.

While I hate to draw examples, could you imagine the borg not detailing in idiots terms the differences between its current bug ridden, code bloated pile of effluence and the next version in terms that your gran could not understand?

If I said to someone on OS 3.1 that the advantages of going to OS 4 were a, b and c and they were geeky sounding, they'll ignore them. But if I said a, b and c in terms of what they could understand (for example, Impression runs twice as fast or the CD will work correctly and enable you to read any form of disc in it or your disc access is four times faster and you're not restricted to small drives any more), then they are more likely to say "yeah - neat!" and buy the thing.

It doesn't matter what they thought then - they were obviously wrong. No easy way to say it, but they were hideously off the mark. And looking at the spec sheets, they still are. Sure, Castle are doing a much better job at describing, in human terms, what the benefits of OS 5 are over the RPC, but it's still probably not enough.

Maybe the pair of them can join forces and employ Chris as some form of media chap - no wait, last time they agreed to do something together....

;-p

Nodoid on 15/1/06 1:04AM
[ Reply | Permalink | Report ]

I herewith nominate Chris for the RISC OS Awards 2006!

hzn on 15/1/06 7:49AM
[ Reply | Permalink | Report ]

Fantastic article! Something like this has been needed for years, and for such a clear, concise article to finally be delivered is a huge step forward for the platform. Well done Drobe!

One thing I would still like to see however is a decent resource comparing the major software packages and their features for new users. Some straightforward reviews with screenshots would be very helpful; it would be good to include older software still available from dealers, especially since many websites for the software have long since vanished.

timephoenix on 15/1/06 7:49AM
[ Reply | Permalink | Report ]

Two very good RISC OS features are missing: 1. Filer operation can be paused and and show progress 2. Scroll direction etc. can be reversed using other mouse button saving lots of mouse movement

hzn on 15/1/06 11:49AM
[ Reply | Permalink | Report ]

Agree with the above, a great article Chris.

Here is another feature I would like: When you drag a file to save (from say, !Zap) into !FTPc, it saves it to the FTP site you are viewing, which is great. However, after that you can't just press F3 then Enter to save, you have to drag the file every single time, which is a bit of a pain. Not sure whose place it is make something like this work, but it would be truly great to give every RISC OS app the ability to save to FTP this easily.

thegman on 15/1/06 1:35PM
[ Reply | Permalink | Report ]

thegman:

FTPFS - what a great idea

Having said that I prefer drag'n'drop - a core RISC OS function

ROHC on 15/1/06 1:53PM
[ Reply | Permalink | Report ]

Oh yeah, and also filesystems for SCP and SFTP would be great. :-)

JGZimmerle on 15/1/06 2:09PM
[ Reply | Permalink | Report ]

To ROHC: FTPFS ist indeed a good idea and if it is a FilingSystem then Drag'n'Drop etc. is there automatically since that's then the filer part..

To ROL/CTL/...: A good FTP-Server as well as a server for Windows shares (that is smbserver in good) would be nice as part of RO too. Hi, ROL, include that in Select and Select for IYONIX and quite a few more user will be seriously tempted to upgrade!

hzn on 15/1/06 2:42PM
[ Reply | Permalink | Report ]

In reply to ROHC, JGZ, hzn,

While FTP FS would be great, !FTPc fits the bill for me perfectly, except for that one feature, I don't really mind having to a seperate app for FTP at all, especially one as good as !FTPc. I think all !FTPc needs to be able to do is to 'pretend' to apps that it is a normal Filer window, so that all apps can open/save files using it. I'm not sure if that is something which needs to be done by ROL, !FTPc authors, or other application authors, but if it could be made to work, all apps would be able to save/open from FTP, which makes a great improvement from say the Mac, in which some apps like BBEdit (a text editor) can save to FTP, but the majority cannot.

thegman on 15/1/06 3:11PM
[ Reply | Permalink | Report ]

A couple of things missing from the RISC OS 5 list:-

DOSFS supports FAT12/16 and FAT32, including Windows broken versions of FAT16 with a FAT32 style variable length root directory (which previous versions including the ones in RO4 & Select does not).

LanmanFS supports proper long filenames with RISC OS filetypes translated in to/from DOS extensions by MIMEmap and ,hex extensions for other RISC OS types. LanmanFS on Select only supports 8.3 DOS name, munches 10 character RISC OS names to 8.~2 and stored types in hidden RISCOS.EA files.

RISC OS 5 supports the scroll wheel of USB mice. Select also has support for the scrollwheel of PS2 mice.

druck on 15/1/06 3:30PM
[ Reply | Permalink | Report ]

Nice piece. One bit seems missing, though - a change summary.

E.g.

Here's what was in RO2. Here's what the final version of RO3 added. (Then comes the tricky bit of defining the split where RO5 left RO4 in its Adjust/Select incarnations behind.) Here's what RO5 adds. Here's what Select/Adjust adds.

As an interested but currently non-Acorn using outsider, I have no idea what Select/Adjust can do that RO5 can't, so it's hard to understand why people want Select/Adjust on Iyonix or why there is any point to making a 32-bit version of RO4 when RO5 already exists.

In the barrage of facts and figures in this article, one cannot readily pick out the difference between current RO4 and current RO5.

Otherwise, it's good.

lproven on 15/1/06 6:08PM
[ Reply | Permalink | Report ]

timephoenix: regarding a comparative list of software titles - is this something that could be combined with the risc-os.net database? In a similar vein, something I've been working on a bit at my site (linked from my user info) is a 'must have' list for RISC OS, detailing essential apps and utilities, and where to find them.

jymbob on 15/1/06 8:08PM
[ Reply | Permalink | Report ]

jymbob:

Integrating such reviews into RISC-OS.net and the like is exactly what I was thinking. I think probably the weak point of such directories is their size - they tend to cover every tiny little app, hence making them easy to get lost in. For new users who just want to know what options there are for web browsing, DTP or whatever, it's too confusing.

My ideal RISC-OS.net: - The current database - Comparisons of current and legacy Machines - Comparisons of major software packages - How-to guides on using RISC OS for certain tasks (Drobe's feature on Antius [link] is a good example of this) - Converting to RISC OS - guide to changing to RO from another platform

For such promotional material to be available on one of the major RO portals would be very positive; at present there really isn't one place to easily find out what RISC OS is all about. Any effort to create a newbies guide like this would certainly gain my assistance.

timephoenix on 16/1/06 6:51AM
[ Reply | Permalink | Report ]

timephoenix:

"Converting to RISC OS" - perhaps one of the welcome guides could be published online.

My wife, who had only really used windows before she met me, was most impressed with the one that came with the Iyonixes.

ROHC on 16/1/06 9:54AM
[ Reply | Permalink | Report ]

Iproven: People want it done for the very good reason of there not being two different current versions of the OS, and that's before you start thinking about features. The current situation is only just about bearable if you assume there are never going to be any more new machines (A9 notwithstanding), and I really hope that there will be.

SimonC on 16/1/06 10:12AM
[ Reply | Permalink | Report ]

SimonC: I really don't understand your comment. It seems to me to be a /non sequitur/ and to not actually understand my point. I am not criticizing here, merely seeking clarification.

Firstly, my overall point was requesting a summary of /differences/ as well as merely a feature list. You don't seem to be responding to that at all.

Secondly, I mentioned parenthetically that I didn't understand why people seemed to want RO4S/A on Iyonix when Iyonix already has RO5. Many of the changes being made to RO4.4 according to this article are playing catch-up to what's already in RO5, as far as I can tell. E.g. 32-bit cleanliness, removal of Acorn device independence, moving of H/W-specific drivers out of the kernel into separate modules, etc.

This seems to be what you're answering, yes? In which case, you seem to be "correcting" me by rephrasing what I said.

There /are/ 2 current versions. Porting RO4S/A onto RO5 machines means perpetuating this instead of merging RO4S/A & RO5. Fragmentation is always a bad thing, I feel, and the RO4-5 split is really harming the already tiny & struggling RISC OS world. Both Castle and RO Ltd are shooting themselves in the foot every day that RO4 & RO5 continue as distinct entities and it is very likely to doom the entire platform.

So why are people so keen that they are pledging money for a version of RO4S/A for Iyonix? What is so special about RO4S/A that people want to cripple the Iyonix further by having 2 subtly incompatible versions of the O/S? That is something I hoped this article might tell me, but sadly, it doesn't, which is why I commented. I didn't want to start a RO4-vs-RO5 flamewar or something.

lproven on 16/1/06 10:41AM
[ Reply | Permalink | Report ]

Iproven - What you seem to be also missing is that Select/Adjust has a lot of features that RO5 is missing. This is what Select for Iyonix would give.

DS1 on 16/1/06 10:47AM
[ Reply | Permalink | Report ]

DS1:

Ain't no such person as "iproven" but I hope I'll do.

I am /not/ *missing* it! I have posted 2 comments *specifically asking* for a summary of the differences? How much more explicit can I be?!

On the other hand, what you seem to be missing is that in a market of perhaps a few thousand users of what is by modern standards a very expensive, very slow and very feature-poor computer, having 2 incompatible variants of the unique, weird, proprietary OS is absolute commercial suicide. The only way for proprietary software & hardware companies to survive is to sell. The only way to consistently keep selling is to get new users. The only way to attract new users is to offer compelling advantages. RISC OS has very few of those. To compound its difficulties by trying to explain that there are 2 versions of the OS and what you can do and what you can run depends on which version you have is absolute commercial *suicide*.

RISC OS is probably doomed anyway, but unless this is sorted and there is one version that runs everywhere, it will die even sooner.

I welcome any reasoned argument as to why this is not the case.

-- Liam P.

lproven on 16/1/06 11:03AM
[ Reply | Permalink | Report ]

lproven - So you didn't write : "Secondly, I mentioned parenthetically that I didn't understand why people seemed to want RO4S/A on Iyonix when Iyonix already has RO5. Many of the changes being made to RO4.4 according to this article are playing catch-up to what's already in RO5, as far as I can tell. E.g. 32-bit cleanliness, removal of Acorn device independence, moving of H/W-specific drivers out of the kernel into separate modules, etc." ?

You claimed you didn't understand why people want Select on the Iyonix. The reason is simple. There are things in RO4.39 (Select/Adjust) that are not in RO5. Just as there are things in RO5 that are not in select. My understanding, subject to clarification by ROL or the Select authors, is that Select on Iyonix will be a number of soft loaded modules adding the extra stuff, without taking away what's already there.

Dave

Dave

DS1 on 16/1/06 11:12AM
[ Reply | Permalink | Report ]

Iproven: You asked why people wanted a RO4.39 or whatever version of Select on the Iyonix, and what the point would be of doing it, and I gave you one answer, which could arguably be a more important one than the difference in features. I wasn't "correcting" you at all. You agree in your later post that the fork is bad, and is that not a good reason to want Select on the Iyonix? To help remove the fork?

I wasn't replying to all of your post (i.e. the bit asking about the differences) because I don't know much about the answer to that one.

SimonC on 16/1/06 11:39AM
[ Reply | Permalink | Report ]

Good article with considerable care and attention paid to RO4.XX and Select features.

I remain to be convinced that Select is the way forward - RO5 *works* and I see nothing in the above articles that would prompt me to want to sign up for it. As to two forks - surely if *everyone* doesn't keep their Select subscriptions up to date there will be a "plethora" of forks (one for each different Select variant various people dropped out at)?

RO5 has additional features (as mentioned by Dave Ruck) and, to be honest, it does what I need in the way I need it. It also works and is available *now*. Additionally I am not convinced I won't lose functionality if I were to purchase Select and it were to simply "trample" over large portions of RO5. Given the relatively low priority ROL seem to give to Iyonix I personally would be happier sticking with Castle. At least RO5 has an open published API and free updates - surely it represents a better path forward than Select for Iyonix users?

AMS on 16/1/06 1:43PM
[ Reply | Permalink | Report ]

Surely what's needed here, and what I believe to be happening, is ROD bringing Select into line with RO5, by augmenting whatever improvements are already in place.

I for one, wouldn't be happy using Select on my Iyonix unless I was certain that everything that worked under RISC OS 5 would still work, and in the way I was expecting it to.

I can't imagine either party would want any of their functionality reduced by a merge.

jymbob on 16/1/06 4:03PM
[ Reply | Permalink | Report ]

Iproven: before you disappear under an avalanche of pro- and anti- Select-for-Iyonix protagonists, have a look at www.iyonix.com/iyonix/features/503notes.shtml which you may find enlightening. A similar analysis exists on the ROL site, as I recall.

bucksboy on 16/1/06 6:56PM
[ 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

Acorn Cybervillage needs new webmaster
Campaign against dead websites

34 comments, latest by quatermass on 18/02/05 11:36AM. Published: 11 Feb 2005

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
"It gives a bad impression to anyone wondering who's going to be next to go when the web sites reporting the news have stopped reporting anything"
Page generated in 0.108 seconds.