Decreasing compatiblitiy of OS4 a good thing?

There is a topic about selling a system where someone said decreasing 68k compatiblity of OS4.1 is a good thing. I was really surprised about that view. Decreasing compatiblity means less software works. For me that is only justified if at the same time new different software is added, f.e. it is easier to get modern ports. A operating system is the base, not more and not less. People use a computer to run software, be it for entertainment or doing serious things.

Decreasing the compatibility IS THE ONLY SOLUTION to move forward. You want to use 60k softfware? use EUAE

I think compatibility with 68k HUNK binaries is mandatory, given how much of the original software library does not have source code available

Software compatibility with 68K software was the main reason I purchased my OS4 system way back in 2004. It allows me to run Final Writer 97, TVPaint, PPaint and ARexx. Some of my other software only works under emulation with E-UAE running on OS4.

So I have a few questions ...

If we get rid of 68K compatibility, where are the software products that will replace our old favourites?

Are there actually programmers willing to do the work?

I am not a developer or a beta tester. I'm just a regular user.

Here is a snapshot of the software I use most on OS4:

NotePad
Multiview
Final Writer 97
IgnitionOS4 spreadsheet
TVPaint
PPaint
AmiDVD
IBrowse Web Browser
NetSurf Web Browser
Odyssey Web Browser
Soliton MUI-based card game
Many ARexx based scripts and programs

---
redfox
MicroA1
AmigaOS 4.1 FE Update 3
Enhancer Software Core Edition

I purchased Enhancer Software V1.5 and Enhancer Software V2.2
but currently have installed only Enhancer Core, AmiDVD,
SmartFileSystem v1.293 and diskcache.library.

Quote:

My reason to buy AmigaNG was need for more power, my last Amiga - A4000T with CyberstormPPC was simply too slow for daily work. In 2003 I purchased Pegasos 2 with MorphOS and in 2013 AmigaOne X1000 with AmigaOS4. I have never regretted it. I used them for daily work many years.
All old great 68k games I am running with E-UAE.
The only 68k application I am using often is NetFS on AmigaOS4. Morphos NetFS is native, so I don't need it here.

Compatibility with 68k application was great in time of MorphOS 1.3, where was no native network stack. But today it is practically not needed.
Of course, compatibility with system friendly 68k app is not bad.
More important for me is to have near-modern web browser, IMAP email client, Excel/Calc and Word/Writer compatible apps. And multicore would be useful. That's enough for it to be my daily work system.

Sorry, I do not see the benefit of having an "Amiga NG" with an "Amiga OS" that does not support 68k programs.

I personally use a lot of 68k programs that have no PPC port:
- CubicIDE
- ArtEffect
- StormC 4
- IMP3
- AmIRC
- VinCED (because CON is seriously buggy)

And no, running them in UAE is not an option for me.

Quote:

You are right, I forgot CubicIDE. I am using it too, but more often I am using Lite-XL. It is not IDE, but is very useful. I also ordered StormC5ED.

@FlynnTheAvatar

Quote:

clearly, we need better debug tools. I’m working on something.

AmigaOS back compatibility is very important for me at least for system friendly apps and tools. In 68k we have lot of apps and tools never replaced by equivalen in PPC like Adpro, ImageFX, Art Effects, WCS, Vista Pro, etc..

Furthermore, I think that compatibility should be improved even further. Some app with PPC support for PowerUp libs not works with actual AmigaOS4.1 like Imagine 3D PPC, Tornado3D, ImageFx PPC modules, etc.. only some WarpOS apps but not old PowerUp.

Classic emulation modes of EHB, HAM, 256c should be updated for works with modern GFX cards ( actually show bad colors ) and add a lowres with doubled pixels be necesary. Updated Petunia with better JIT are welcome too.

I not want a new system without compatibility with old 68k apps. I love contiue using my old Amiga apps in my AmigaNG at fast speed of Petunia.

Probably what it mean originaly and what mean Andrea is to ditch 68k from OS internals and co, but of course, keeping some sort of emulation/sandboxing integrated well or whatever. No one saying that 68k should be no more runs on os4, but not in a way it happens now internally. Just internally it should be different, and visually the same as now.

But, we all know this probably will not happen because of too much work need it to make it all allright

If I understand you correctly, you mean ditching old 68k stuff from the APIs/libraries. IMHO this is even worse, because this makes maintaining software for all Amiga flavors (3.x, MorphOS, AmigaOS 4.x) much more compliate than it needs to. And it is pretty bad with the interfaces for libraries calls already.

@FlynnTheAvatar
I mean to remove everything 68k from os4 internals, yes. API can be left as it, just added new os4 specific (as it happens right now), but to ditch everything from internals, but sandbox 68k properly so everything will runs like it now, and users will not notice. And sure there mean just only os4 apps, without matain os3/morphos/aros with same code, it always was a mess with big projects, and .. and it all same old stories :) That just can't be done for today properly with current developer resources.

Main point that for users it still all of course should be same, just internaly a bit more sandboxing. But as i say this all just cheap talks, because in reality it all will be like this is now, with some new things added on top and bug fixing. Such radical changes probably will have no place ever.

Please forgive this nostalgic journey ...

My first Amiga was an A2000HD with Workbench 1.3 and a generous suite of software. I upgraded to AmigaOS 3.1 and added some more RAM, CD-ROM drive and flicker-fixer card. I did not have an accelerator card, so all my software was compatible with 68000 CPU, including Final Writer 97.

My next Amiga was my MicroA1 with AmigaOS 4.0 Developer Pre-release software. I purchased a motherboard bundled with OS4 and received the motherboard and a white CD-R containing OS4. I think it was the first update that officially supported the MicroA1. After many enquiries and many months of waiting, I finally received my official AmigaOS 4.0 Pre-release CD and I was able to register with Hyperion and receive the next software update.

By that time, I had been using OS4 for many months and I was transfering software from my old A2000HD to my MicroA1. Most system friendly software worked just fine with OS4, however, I needed to use E-UAE for some of my old favourites (Appetizer suite, Microfiche filer and Amiga Vision).

I was also collecting some other 68K software, such as TVPaint, PPaint, Imagine and Real3D.

As the years passsed, more PPC compatible software became available and I now have a mixture of 68K and PPC software.

-----

So, back to the 68K compatibility issue ...

If it is so important to get rid of 68K compatibility, why not just bite the bullet and produce your new OS with no compatibility whatsoever. Call it XYZLinux with a nostalgic "Amiga" skin on the desktop. While you're at it, migrate to some other processor and let PPC and 68K rest in peace. Just let us know ahead of time so we can decide if we want to join your journey.

---
redfox

If I'll lost 68k compatibility layer I'll lost 60% of software I use with my NG system.
If OS4 lost 68K compatibility and left posibility of use older apps exclusive to use of UAE It's not very diferent to use Linux, Mac or PC with UAE.

As Kas1e say, really we no need use Petunia an compatibility layer of libs if a new sandbox can do the work and it's transparent or invisible to user and corrected integrated in the system can be done too.

I think I agree with the majority of folks here. If we lose 68K support, we will lose part of the Magic of OS 4. For my money, I would push harder to have better integration and support for older software.

As stated, a sandbox for 68k OS friendly software is the way to go.
All legacy components should be removed from the OS4 source code.

Another possible path could be to build a spin-off of the current OS4 state, redesigned with cleanups and architectural improvements to make the best possible use of modern hardware.

My two cents.

@afxgroup I think you're alone now

I don't care.. All other OSes have moved from roots using sandboxes and something like that. Only OS4 doesn't want to grow and stay rooted to the past because 3-4 software people is using and there are no alternatives. So you have an ancient OS, with ancient software and no new users, no new developers.. But hey! You can use ArtEffect that modern people can use in a browser...

@NinjaCyborg
He is not.

@All
I think most of you don't get properly what mean "ditch legacy support". It didn't mean _remove_ ability to run transparently 68k apps ! Of course not , we all there with roots from 68k all in all.

What it mean, to ditch support of 68k legacy _internally_ in the OS itself, and instead, make 68k stuff be sandboxed/handled properly: i.e. it will be the same as you have it now, you just run it, and petunia/uae/whatever else, deal with it and you think like it the same as before, maybe with some small diffs. But _internally_ for the OS4 it will be just running emulator (as you have it now anyway with Petunia already), just, with ditching old non need it code in the core of the OS4 itself.

Do not get what it all fuzz about.. Even if it will be done some day ever, it will be just the same as now : you run 68k binary and it works and share same os4 windowses, etc,etc , just like it was done long ago for example on AROS. It even will be better, as banging hardware stuff will works better.

And while it should be done to move forward, i fear it will not because of limited developer resource. But even if it will, users will have no issues and will not notice.

Just some of you heard "ditching 68k", and "ahhhha, they want to remove ability to run old stuff!!!" , no, it mean ditching internally, so os will not rely anymore on ancient code, and will handle and sandbox things properly, not just like Petunia now, when you can't use banged software and when developer need to know that there A0 and D1 registers when code on C for OS4, which only make things be harder.

It is the old Amiga users who insist on 68k compatibility and do not want to allow any changes. Personally, I still see myself as a newcomer to AmigaOS 4.x, even though I already owned PPC hardware (AmigaOneXe) 20 years ago.

At that time (AmigaOS 4.0pre), it was unthinkable to rely solely on native software because there were virtually no alternatives. Things have changed in the meantime, and there are native alternatives for just about everything.

Of course, it's great that we still have the option to run 68k software through Petunia or E-Uae. But if this means hindering the growth of AmigaOS 4.1, then I am against this ancient 68k layer. By Amiga standards, we already have powerful X1000/X5000 hardware where everything can be emulated cleanly and quickly.

As already mentioned, if it should still be possible to use 68k software in the future via a sandbox or other means, but the user notices as little of this as possible, then this should be done. Everything, and I mean everything, that hinders the further development of AmigaOS 4.1 should be dropped.

I know the comparison is a bit off, but Apple did the same thing with MacOS when switching from PPC to Intel and ARM. However, with the next major operating system versions, this support will be dropped in the future.

Of course, this is all wishful thinking and none of it will happen, but I still wanted to share my thoughts on the matter from a user's perspective.