Aemulor - 26bit emulator for Raspberry Pi 'PreOrders

[Davespice]

Looks like it's not going to be free;
http://www.cjemicros.co.uk/micros/indiv ... EM-AEM-RPI

Obviously the guy who has done the work deserves to be rewarded, but I think the cost will likely put off the vast majority of people who would like to use it for the fun of playing Archimedes games like Elite.

Please don’t launch into a massive rant at me folks. It’s a fairly obvious point.

I used to live in Worthing too... maybe I've met him in a pub... *racks brains*

[polas]

Cool, I suppose paying for it was to be expected and £18 isn't a huge amount. However, I am currently using ArcEm on my RPi which is a very satisfactory solution and provides, generally, good backwards compatibility. Does anyone know what advantages Aemulor will have, over and above ArcEm, in order to tempt people to part with their cash?

Nick

[Davespice]

As far as I know it allows you to run old programs directly on the Pi without having to Emulate an entire Archimedes desktop etc. It just fixes the 26 bit to 32 bit difference and a lot of the RISC code can then be executed semi-native by the ARM. So I presume the main advantage is speed?

Please correct me if this is wrong.

[EastLight]

Looks like it's not going to be free;

It also looks like each license will be be tied to a single Pi via the Ethernet MAC address (although I'm unsure as to how the advertised Model A compatibility will work if this is so).
Such a policy would prevent SD card swapping between multiple Pis.

[DexOS]

Coders are no different from other people, they need paying.
The reason kids do not want to code any more, is because they know most of its free.
People need to see software as having value, the only way for this to happen is charge for it.

[AMcS]

Coders are no different from other people, they need paying.
The reason kids do not want to code any more, is because they know most of its free.
People need to see software as having value, the only way for this to happen is charge for it.

Agreed, part of the problem is that frequently the efforts of programmers (and the skills involved in producing code) are themselves devalued if the code produced is made available freely or (at most) for a small nominal fee.

I always found it odd that people would have no difficulty paying large sums of money for some piece of computer hardware and then have a problem paying a fraction of that cost to acquire software that actually makes the hardware functional and worth having.

[EastLight]

Coders are no different from other people, they need paying.
The reason kids do not want to code any more, is because they know most of its free.
People need to see software as having value, the only way for this to happen is charge for it.

Agreed, part of the problem is that frequently the efforts of programmers (and the skills involved in producing code) are themselves devalued if the code produced is made available freely or (at most) for a small nominal fee.

I always found it odd that people would have no difficulty paying large sums of money for some piece of computer hardware and then have a problem paying a fraction of that cost to acquire software that actually makes the hardware functional and worth having.

So in that case, all freeware should be banned, and all commercial software sold subject to draconian, 'one-license-per-machine' digital rights management, correct?

[DavidS]

Coders are no different from other people, they need paying.
The reason kids do not want to code any more, is because they know most of its free.
People need to see software as having value, the only way for this to happen is charge for it.

Agreed, part of the problem is that frequently the efforts of programmers (and the skills involved in producing code) are themselves devalued if the code produced is made available freely or (at most) for a small nominal fee.

I always found it odd that people would have no difficulty paying large sums of money for some piece of computer hardware and then have a problem paying a fraction of that cost to acquire software that actually makes the hardware functional and worth having.

So in that case, all freeware should be banned, and all commercial software sold subject to draconian, 'one-license-per-machine' digital rights management, correct?

No, I would say that it is the developers decision. The trouble is that the Open Source and FreeWare tends to be prefered by users even if there are far better commercial products. I do think that some comercial software is a bit over priced, though that is just my view.

I think that comercial software should be promoted well. Free Software still has its place. Commercial software should be priced in such a way as to cover the development costs and keep the developers paid in a reasonable way. It has been a while since I had a comercial software release, though my rule is to price things based on how wellall current products are selling, and ensuring that I maintain operating costs, while providing an income of $1000 per month per employ. (my current income is less than half of this). Of course the employers tend to look at things differently, most of them are paying there developers greater than $8000 per month (and I cn understand this as there are surely some out there that have to support themself and 12 other people on there income alone).

[Muzer]

RISC OS things generally cost lots of money. It's this lack of community/sharing that puts me off in favour of Linux, since I don't have money. For hobby OSes like this*, people should program because they find it fun rather than for money.

If you're just interested in messing about with RISC OS, which realistically everyone likely to use it nowadays is, you're not going to want to spend much on it.


*Come on, you don't seriously expect RISC OS to be taken as a serious OS in this day and age with its lack of features.

[DavidS]

RISC OS things generally cost lots of money. It's this lack of community/sharing that puts me off in favour of Linux, since I don't have money. For hobby OSes like this*, people should program because they find it fun rather than for money.

If you're just interested in messing about with RISC OS, which realistically everyone likely to use it nowadays is, you're not going to want to spend much on it.


*Come on, you don't seriously expect RISC OS to be taken as a serious OS in this day and age with its lack of features.

WHAT LACK OF FEATURES?
The only two features that I can think of that RISC OS does not have are Multithrading, and Disk Paged Virtual memory.
We do have the availability of Preemptive multasking for programs that just can notbe made to work with out it. And we have every other feature that makes an OS "Modern", there are even those of us the are experimenting with adding multiprocessor support (eg SMP) to the OS with outloosing the helpfulness of coooperative Multitasking.

Just about all of the complaints that I have heard regarding percieved lack of features have nothing to do with the OS, though rather available aplications and device drivers. To correct this we need software developers. Most developers wil donate there time to create some things, though other things are just to time consuming to expect them to do in total for free.

I do some things free, and there are some things that require to much time, and as such require some form of reembursement. I enjoy software development, if the cost of living were 0, zero, zip, then I would gladly do everything for free, though there is the cost of Food, Taxes, Electricity, Internet service, Gas, Clothing, Electronics components, Computers, test equement, GOOD Compilers, Soap, Water, etc...

As to the two thigs that I sited abve as lacking from RISC OS:

Multithreading: This on should probably be addressed, though for the time being it is possible to work around it fairly easily (actually easier than using multithreading in a n*x OS).

Disk Swapped Virtual Memory: There is no need for it, all that it generally does is slow things down, and needlesly swap out pages while there is plenty of pysical memory not beeing used by anything. I could see this in the days of computers having less than 512KB of memory, though today we are using a computer that has 1024 times this (512MB =512KB * 1024).

[AMcS]

Coders are no different from other people, they need paying.
The reason kids do not want to code any more, is because they know most of its free.
People need to see software as having value, the only way for this to happen is charge for it.

Agreed, part of the problem is that frequently the efforts of programmers (and the skills involved in producing code) are themselves devalued if the code produced is made available freely or (at most) for a small nominal fee.

I always found it odd that people would have no difficulty paying large sums of money for some piece of computer hardware and then have a problem paying a fraction of that cost to acquire software that actually makes the hardware functional and worth having.

So in that case, all freeware should be banned, and all commercial software sold subject to draconian, 'one-license-per-machine' digital rights management, correct?

Don't recall saying that. It's up to the software author as to what they charge, if they overcharge they won't sell stuff.

If they're a commercial outfit and get draconian people will avoid using their products.

I have no problem with authors choosing to release freeware if they wish - the only issue is that if the public view software as "without value" (sure it's free - so can't be worth much...).

Would someone building computer hardware work for nothing ? No, so I would not expect a programmer to do so either (if they choose to fine - hopefully they can find some other way to pay their bills, pay for a roof over their head and for food)

[AMcS]

RISC OS things generally cost lots of money. It's this lack of community/sharing that puts me off in favour of Linux, since I don't have money. For hobby OSes like this*, people should program because they find it fun rather than for money.

Most of the software available for RISC OS is *free* (as in without charge). If you're using RISC OS itself on Pi you're doing it for free. If you look at the discussion over on the SWI documentation page you'd see many FREE IDEs being suggested - none of these cost a penny. You get the PRMs (programmers' reference manuals - albeit in PDF format - for *free* on the RISC OS Pi distribution). There's plenty of "sharing" going on there if you're prepared to notice it.

If you're just interested in messing about with RISC OS, which realistically everyone likely to use it nowadays is, you're not going to want to spend much on it.

And does anyone seriously use Linux a multi-user, resource hungry OS on a board with a CPU that has (around) half the power of a pretty "oh..hum" SmartPhone? Yep guess what they're messing too... that's the whole point of Raspberry Pi - to investigate - to try out to (for want of a better word) MESS.

There is also arguably a community working on RISC OS to improve it - have a look over at https://www.riscosopen.org

*Come on, you don't seriously expect RISC OS to be taken as a serious OS in this day and age with its lack of features.

Why not? The features are software that can be added, RISC OS being highly modular can be extended in a "transparent" manner. (Sure RISC OS is doing stuff on Pi it couldn't do before - it couldn't read SD cards a few months ago and now it can - for one).

RISC OS in the past ran on the fastest desktop computers you could buy - it was the first OS with inbuilt fully antialiased scalable fonts, it could boot in seconds because it was in ROM. It has a lot of features that are desirable on embedded and portable devices - exactly where computing is going. And to cap it is optimised for and runs on the CPU architecture that is the most common on the planet (there are MORE ARMs on Earth than people !).

So yes RISC OS has a future and I would not write it off so quickly.

[Muzer]

RISC OS is a lovely OS in its ways, don't get me wrong, and I do thoroughly enjoy playing around with it, and it does have some lovely innovations that it's a shame other OSes have never picked up. I'm not denying that it's a good OS to mess around with and have thoroughly recommend it to friends if only to broaden minds a bit. But I think it's probably got to the stage now where it's hard to use it for general desktop use, due to the lack of application software, especially compared to, say, Linux, and I would never expect it to once again reach its former greatness.

As for missing features, I was mostly thinking pre-emptitive multitasking, which adds reliability and probably slightly improves performance too - it's quite important that an app would never be able to crash the system, and I don't think you can ever guarantee this in a co-operatively multitasked OS (please prove me wrong!). I didn't realise it didn't support swapping to disk but that's a pretty important feature for if it ever gets some large (some would say bloated) piece of application software that doesn't quite fit in RAM . Also lacking are a decent level of security for a modern OS (just as one example in the usefulness of this, in this age of internet access it's nice to be able to run an untrusted application knowing it's not able to modify any important system files, for instance - you can do this on Linux and you can just about do it on Windows), and at the more mundane level, lack of support for modern features such as (two off the top of my head) drive partitioning and wifi. There are probably many more OS features that I've left out here, but you get my point. I'm not saying that all of these are unsurpassable, but they'd certainly require a lot of work. I'm not saying that nobody should be doing this work - it'd certainly be a fun thing to do as a project - but my point is that RISC OS, unless it suddenly receives significantly more investment than it's recently had, will remain a niche interest OS. It's sad, because as I said, it is a wonderful OS, but IMHO it's the truth.

[DexOS]

its is a great OS and if people would value programmers and pay them for the software, it could be brought up to date.
People think nothing about paying £2.75 for a cappuccino, but moan if a app is not free or less than a pound.
We need something like kickstarter for software, so a group of people can pay coders to code up to date software, for this and other OS's like it.
This may help kids to make money in there spare time, who have invested the time needed to learn to program.

[DavidS]

RISC OS is a lovely OS in its ways, don't get me wrong, and I do thoroughly enjoy playing around with it, and it does have some lovely innovations that it's a shame other OSes have never picked up. I'm not denying that it's a good OS to mess around with and have thoroughly recommend it to friends if only to broaden minds a bit. But I think it's probably got to the stage now where it's hard to use it for general desktop use, due to the lack of application software, especially compared to, say, Linux, and I would never expect it to once again reach its former greatness.

I see your evaluation. It may be behind in the area of available applications, though not as far behind as you make it sound. This is not an issue of the OS it is an issue of getting the Applications in there.

As for missing features, I was mostly thinking pre-emptitive multitasking, which adds reliability and probably slightly improves performance too - it's quite important that an app would never be able to crash the system, and I don't think you can ever guarantee this in a co-operatively multitasked OS (please prove me wrong!).

I would further add that it is quite important for a user not to ever run 'Untrusted cod on there machine. Yes it is possible to get around the security provided by Linux (tere are a lot of proofs of this just take the time to look around the net), so untrusted code is equaly bad regardless of the platform on which it runs.

Now as to the issue of preventing one application from taking over the whole system, this comes down to adding tool to preempt a WIMP Task if it executes for more than X centaseconds, this is an improvement that I have felt is needed for some time. One of the advantages of RISC OS is that it is still possible to single task non WIMP Tasks.

Now if you want Preemptive multitasking for apps that require it use WIMP2. Though I would strongly argue that it is possible to get better performance out of well written cooperatve multitasking code, as the time spent runing the lists in the scheduler of a preemptive multitasking system is quite significant.

I didn't realise it didn't support swapping to disk but that's a pretty important feature for if it ever gets some large (some would say bloated) piece of application software that doesn't quite fit in RAM .

Find a good reliable Swapping algoritm that does not ever swap out to disk unless it is absoutely needed and then I will agree, until then the current philosophy of swapping the LRU on the second pass of the clock algorithm regardles of how much physical memory is unused by anything will continue to slow things down (Windows is realy bad about this, linux is a lot better though still far from ideal, the best I have seen is that used in Mac OS 7.5.5 through Mac OS 8.1, (they canged it in the urealeased 8.2 and kept the change in 8.5 just to changeit even more in 8.6).

Also lacking are a decent level of security for a modern OS (just as one example in the usefulness of this, in this age of internet access it's nice to be able to run an untrusted application knowing it's not able to modify any important system files, for instance - you can do this on Linux and you can just about do it on Windows), and at the more mundane level, lack of support for modern features such as (two off the top of my head) drive partitioning and wifi.

Drive partitioning is the doain of the FS (think driver) not the OS in RISC OS (As well as most other modern OSes we are not using DOS, CP/M, RTX, etc anymore). Wifi once again a driver issue that has nothing to do with the OS, if you want it before I get the USB stack up to date and I add a WIFI MODULE myself then WRITE YOUR OWN WIFI MODULE.

So many people have complained about it, ando none of them have wanted to do the work, so I decided that as soon as I am done updating the ROOL USB stack I will implement support for Wifi networking devices using the Atheros chipset. I can not say for sure how long this will take me, if you are in a hurry to have WRITE YOUR OWN WIFI MODULE.

As to your security concerns do not run an application, module, etc on your RISC OS computer ifyou do not trust the developer of the software. This is not windows, nothing is going to caus it to run from a web script or a buffer over run in a picture decoder (the stuff that windows just got wrong).

There are probably many more OS features that I've left out here, but you get my point. I'm not saying that all of these are unsurpassable, but they'd certainly require a lot of work. I'm not saying that nobody should be doing this work - it'd certainly be a fun thing to do as a project - but my point is that RISC OS, unless it suddenly receives significantly more investment than it's recently had, will remain a niche interest OS. It's sad, because as I said, it is a wonderful OS, but IMHO it's the truth.

It seems that you think that less is bein done for RISC OS at this time than is. There is work on multiuser support, and per user permisions being done that would take care of your concerns about security, the USB stac is being updated trust me it is definitely being updated, I am begging to dream in packets, and USB timeslices. Also before I started working on the USB I had started a port of he Dillo web browser, that I intend to finish once I get done with the ROOL stuff.

[pygmy_giant]

In simple terms does that mean that RISC OS wil poll the USB more intelligently and less frequently, leading to improved performance?

I am already stisfied with RISC OS, so even better would be, well, even better.

The only downside to RISC OS in my opinion is the comparatively low amount of free software available compared to Linux, but if you only use it to hobby program like I do then thats no biggie.

There are bounties on the ROOL website, maybe some kind of further monetising strategy could inject a bit more life into the RISC OS development/software base. Maybe some kind of 'app store' could do that? The incorporation of pledgey pre-ordering might help also.

Personally, I don't mind paying for things if I really need/want them and they work.

Examples of what I could see myself paying for on RISC OS would be a mobile broadband app and a decent 3d racing game for my son.

I bet a lot of open source Linux stuff could be ported with minor alteration by people who know what they are doing.

[DavidS]

In simple terms does that mean that RISC OS wil poll the USB more intelligently and less frequently, leading to improved performance?

Yes it will be much smarter about managing the USB devices. It will also implement the features needed to suppot a much broader range of Hardware, and have a reasonable level of documentation on the implementation.

I am already stisfied with RISC OS, so even better would be, well, even better.

The only downside to RISC OS in my opinion is the comparatively low amount of free software available compared to Linux, but if you only use it to hobby program then thats no biggie.

Well no system on earth has a much free software as Linux, so I guess that we are still ahead of the game on that one LOL.

[Tide]

Examples of what I could see myself paying for on RISC OS would be a mobile broadband app and a decent 3d racing game for my son.

Can't help you with the latter but there's !CommCentre for working with 3G dongles. Guess what: it's free!

[pygmy_giant]

I have it but cant get it working

[jojopi]

Disk Swapped Virtual Memory: There is no need for it, all that it generally does is slow things down, and needlesly swap out pages while there is plenty of pysical memory not beeing used by anything.

Everyone knows that swapping to disk is painful, especially on a Pi. So the fact that RISC OS has a small GUI instead of layers of abstraction piled onto X is an advantage.

As an OS design feature, however, paged virtual memory is tremendously useful. It allows tricks such as transparently using free memory as disk cache or filesystem, mapping files into the address space, demand loading, sharing pages between processes when they are read-only or have not yet been written. How does RISC OS handle this?

[DavidS]

Disk Swapped Virtual Memory: There is no need for it, all that it generally does is slow things down, and needlesly swap out pages while there is plenty of pysical memory not beeing used by anything.

Everyone knows that swapping to disk is painful, especially on a Pi. So the fact that RISC OS has a small GUI instead of layers of abstraction piled onto X is an advantage.

As an OS design feature, however, paged virtual memory is tremendously useful. It allows tricks such as transparently using free memory as disk cache or filesystem, mapping files into the address space, demand loading, sharing pages between processes when they are read-only or have not yet been written. How does RISC OS handle this?

While admittedly RISC OS does not have the best disk caching scheme, I have written a number of OS kernels that do implement standard Disk Paging algorithme (for the practice of theory), as well as ones that do not, and I can say that it is very much simpler to implement a good transparent disk/file cache using free memory on systems that do not implement disk swapping in there virual memory system.

So you did bring up something that needs improved in RISC OS, File read/write caching. Interesting, as the way it is done now cache is kept synced with the on disk copy so often that it is more resilient to a hard power down in the middle of something, or a hard crash, as it is very unlikely to happen when there is anything that is not up to date in the on disk copy.

We realy need a new term for the on disk file that is as easy to comunicate, as many RPi users are using Mass Storage devices that do not have any form of disk to store there "on disk files".

[Mobsie]

Hi,

i know a lot people who switch from Linux to RiscOS now.

All love the Pi with RiscOS and all say is really better than Linux with all the distros and stuppid GUI systems.

Only my words and what i see, nothing against Linux.

Mobsie

[DavidS]

its is a great OS and if people would value programmers and pay them for the software, it could be brought up to date.
People think nothing about paying £2.75 for a cappuccino, but moan if a app is not free or less than a pound.
We need something like kickstarter for software, so a group of people can pay coders to code up to date software, for this and other OS's like it.
This may help kids to make money in there spare time, who have invested the time needed to learn to program.

For ROOL RISC OS there is such a system in place by way of there bounty scheme. See:
https://www.riscosopen.org/bounty/

[AMcS]

Hi,

i know a lot people who switch from Linux to RiscOS now.

RISC OS has an appeal alright, but there's no conceivable reason why people can't switch between the two choosing the more appropriate OS in a given situation - sometimes RISC OS other times Linux.

All love the Pi with RiscOS and all say is really better than Linux with all the distros and stuppid GUI systems.

It's horses for courses, Linux is a top class server OS - it's solid but somewhat resource hungry. RISC OS is written in ARM machine code, nimble, fast and light - on a relatively low powered version of the ARM chip it's going to fly.

On a single user ARM system RISC OS will have the performance edge. On a server though RISC OS wouldn't cut it - Linux would on the other hand be the OS of choice.

New users are, for the first time, getting to try RISC OS - and seeing it on a big HD screen and it does looks the business. Many like what they see and that can only be good for the platform.

[DexOS]

For ROOL RISC OS there is such a system in place by way of there bounty scheme. See:
https://www.riscosopen.org/bounty/

That's great, we need more of this .

The RISC OS should be my ideal OS, as i like both assembly and basic.
But i just do not like the style of the assembly code in the OS, it very hard to trace flow.
It reminds me of the tasm type assembly (yes i know that was x86), with all its red tape.
Plus its GUI, is so ugly, it should go around with a bag on its head (joke)

I have been striping linux down to its core and theres a very nice, fast OS hiding in there .