Maintain Software Compatibility.
All information herein is the view of the author. Use at your own risk, no warenty of any kind is provided.


There is a repeating very bad theme, since about the year 2000, in the release of new OS versions. This theme is the tendancy say that old software does not matter, only currently maintained software needs to move forward so we can sacrifice compatibility. They forget that there is a lot of old software that is very good, some heavily prefered over the newer alternitives.

There is a huge library of very good older software that is still heavily used, do to preferences or percieved better quality (many times likely true to be better) or ability. Thus it is that disposing of the old is a bad thing, as much as reasonable compatibility shoule be maintained. There are many OSes for which this issue is true already, and there are those that are looking this direction for future RISC OS versions. Keep as much compatibility as reasonable, PLEASE

  • A Look at Others Mistakes.
  • Where we are at present.
  • Please Remember the Good.

A Look at Others Mistakes:

Others have released upgrades to there OSes neglecting to maintain compatibility with existing software. This almost universally hurt them, even the ones that later recovered in a way. There is just to much software out there that is no longer maintained and does something in a prefered way, or in some cases for which there is no replacing possible for some reason or another.

When Apple merged with Next to create an updated Operating System there were many Macintosh users that were interested in the possibility. Many of them ended up staying with Mac OS 8.x or 9.x after they tried the new Mac OS X, as the emulator included for running older software was not as good as the real thing. And this was one that at the time did provide a means to run older software, though not natively and that was enough to lose them many customers. Those that stuck with Mac OS X as long as there was support for older software were shaken again when Tiger reached end of life, and there was no longer a means standard to the OS to run older software, many more jumped ship then. Since Apple has done well, though mostly in non-desktop computing (media players, phones and the like). It seems likley they would be in a better position in the computer market had they not pushed out so many long time customers with these choices. It is true that there are the Apple Die Hards, and those that are new to Apple since Mac OS X (now called macOS) was around. How much better would they have done had they kept the loyal classic customer base?

There is a lot of debate in the Amiga world about the well known compatibility issues. I do not know how much this has an effect on the user base, as we still have compatible options available to us. Though it feels like most of us Amigians do not have much interest in MorphOS as anything more than a novelty, and prefer the WarpOS use of hunk type executables over the ELF of AROS, Amiga OS 4 (which still supports the PowerPC Hunk), etc. There are the AROS users, mostly do to the lack of being able to afford newer post-Amiga computers that natively support Amiga OS 4 and thus PowerPC Hunk files (WarpOS is not updated or available for newer HW unfortunately), or the even newer Vampire for the 68K purists.

In 2008 M$ chose to disconue support for Windows 3.xx completely. Then it was not long before we began seeing 64-bit versions of Windows NT (starting with some releases of Windows XP), these no longer have the NTVDI or WoW, so no more Windows 3.xx programs, no more DOS programs, at least not without using a third party emulation layer. Do to the force many continue with M$ NT, though many of them complain about the less than ideal means of running older programs written for Windows 3.xx or DOS. Many of them have become users of other OSes (mostly Linux distros), though many still stay with M$ and complain. This may seem minor when you look at how long it has been since Windows 3.xx was the main Windows version, though people still developed Windows 3.xx programs well past the year 2000, and many of the Windows 3 programs have no newer equalivelent. This issue could have easily been solved by including a CPU emulation layer for NT 64-bit versions, like they do on non-Intel CPUs already.

There are many more examples, though these are the most widely known of them. Even RISC OS had to take steps to remain backwards compatible with 26-bit R15 programs when going to the ARMv5 that no longer supports 26-bit R15 CPU modes, it was one developer that licensed to Castle an effective solution for the Iyonix (Aemulor). Later systems still have the option of updated versions of Aemulor from the author. Without the compatibility of old software it seems unlikely us RISC OS users would still be quite as strong in number. Thankfully RISC OS is still strong, and many programs from the days of RISC OS 2 still run on the newest RISC OS computers running RISC OS 5.26 or newer. For a time there were a couple of other emulation layers for running older programs on newer RISC OS, though the best one is the one that remains.

Where we are at present:

RISC OS still has decent compatibility with older software. This despite a transistion to 32-bit only R15 ARM CPUs. While the usability of RISC OS says a lot about the system as a whole, there is still a very large amount of older software that RISC OS users use on a daily basis that is a big part of the usability of RISC OS as a modern OS.

There are those within the RISC OS world that are advocating that RISC OS move away from its low level API, and to something that would sacrifice compatibility of hundreds of thousands of programs, a large number of modules, and much more. This move would be a desaster for our great OS. Keep RISC OS working, do not destroy it. Some argue that we need to go 64-bit, and do so by breaking compatibility, there is no need for 64-bit on RISC OS yet, we do not have a need for more than 32TeraBytes of dirrect address space (which is what can be reached with 32-bit ARM CPUs depending on the MMU used). And there is not a good reason to leave the 32-bit ARM ISA ever, it is strong, and it will always be around (despite what the crises thinkers preach).

Please Remember the Good:

We have a good OS, the first widely available OS for the ARM CPU, we have a huge amoung of great software for this good OS, we need to keep what we have going for us. We also have a decent base of users writing software for our OS, som of this new software is able to run on everything from an ARMv2 Archimedes on up to the newest ARMv8 RISC OS Systems in AARCH32 mode, and do so quite well at that. We have great music composition applications, we have some of the best Word Processors and other Office Applicaitons, we have graphics Drawing programs that will blow away almost anything else on any system.

Is it really worth losing the good to chase a minimally supported future. Or should we take advantage of what we have, and move to a well supported future?