OlafS25 Quote:

Yes but user experience is in first line desktop. You can perfectly do something with the look & feel but sitting on top of something else. Also you could create layers that hide the different stuff (like linux) and make supporting or transition easier. A little like Apple did when they changed ISA to X86. I see no sense in the try to create another "modern platform" by reinventing all wheels. Besides we do not have the money for that.

The Amiga is a successful chameleon. The Workbench desktop is not always what is seen first. Many games booted straight into the game and the CDTV and CD32 had special menus when not booting into other software. Some users have set up game systems that boot to a scrollable menu of games for easier access with a joystick/joypad. Some Amiga embedded software boots into a custom UI that may not be recognizable as an Amiga (kiosks for example). The Amiga 3000UX booted into Linux but a boot menu is available, as it is on most later Amigas, allowing boot into an AmigaOS partition. The Amiga has support for many input devices too including mouse, trackball, touch screens, light pens, drawing/sketch/art touch pads, joystick/joypad, paddles, etc. The Amiga can look like practically anything with input from practically anything. It can do this with tiny performance and memory by today's standards. It is easy to use and develop for. All this is valuable but it isn't used anywhere today because the hardware is too expensive.

MorphOS adopted the Quark microkernel based on the L4 microkernel. QNX was nearly used to modernize or replace the AmigaOS. Gateway decided to replace much of the AmigaOS with Linux. Amiga Inc. tried to replace the AmigaOS with the AmigaDE (Amiga Digital Environment) virtual machine. ExecSG is a more organic development planned for AmigaOS 4. The AmigaOS needs some serious work under the hood to modernize it but what was extremely unpopular with Amiga fans was Gateway's decision to replace the AmigaOS with Linux and Amiga Inc.'s decision to replace the AmigaOS with AmigaDE. These radical choices solve the lack of modern drivers problem but a return to standard hardware greatly reduces the number of drivers required. I believe organic development and a return to standardized "custom" hardware would be a much more popular decision for Amiga fans than a replacement job or major parts under the hood. One of the major reasons why PPC AmigaNOne failed was because replacing the 68k with a PPC was like replacing an older but popular engine with a more modern but less popular one in a classic sports car. Classic sports cars with engine transplants are less valuable because they are seen as bastardizations. Rebuild the engine, replace internal parts with more modern higher tech parts, add a better flowing head and exhaust manifold, replace the old 8 bit EFI ECU and these tasteful mods are generally viewed favorably by fans. The AmigaOS is no different than the hardware. Most Amiga fans want an upgraded but compatible 68k Amiga and retro is hot right now. This may limit what is possible for the 68k Amiga but the 68k Amiga still has serious potential due to the choices for a dynamic and expandable PC by Jay Miner. Be careful not to throw out the 68k Amiga baby with the bathwater in pursuit of a modern Amiga.

Last edited by matthey on 28-Apr-2024 at 04:22 PM.

@AMIGASYSTEM

Quote:

Also the command c:lha >ram:output.txt *>>ram:output.txt, works fine on AROS, but it generates an empty file !

Can you imagine why?
Instead of directing both stdout and stdert each to same file, you should redirect stderr to stdout and then stdout to the file.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla
Hi, attention, this happened with "one" version of LHA, then I tested other version of LHA and the file "output.txt" is saved correctly with data in it !

@matthey
I don't know if OS4 or MOS can do it, but even AROS x86/68k can start an application or Game "Without starting the Workbench first"

Just start AROS from "Aros Early Startup", and from the prompt run Games and many Applications

_________________
AROS One Home Site
AfA One Amiga OS 3.9

@Hypex

Quote:

RequestFile >ENV:Drawer DRAWERSONLY Set Path "`Echo $Drawer`"

What’s this trying to accomplish?

Set PATH `RequestFile DRAWERSONLY`

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@AMIGASYSTEM

Quote:

AMIGASYSTEM wrote: @kolla Hi, attention, this happened with "one" version of LHA, then I tested other version of LHA and the file "output.txt" is saved correctly with data in

Right. Can you imagine why?

What happens if you redirect stdout to one file and stderr to some other file?

Last edited by kolla on 28-Apr-2024 at 08:18 PM.
Last edited by kolla on 28-Apr-2024 at 08:16 PM.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla
I don't know, if you tell me I will thank you, I am not a developer ! only Amiga and AROS enthusiast I have 70 years, AROS One x86/68k and AfA One are my Distributions !

In the past we discussed about EAB, in many cases I agreed with you !

Last edited by AMIGASYSTEM on 28-Apr-2024 at 08:30 PM.

_________________
AROS One Home Site
AfA One Amiga OS 3.9

Double post, can be deleted

Last edited by AMIGASYSTEM on 28-Apr-2024 at 08:32 PM.

_________________
AROS One Home Site
AfA One Amiga OS 3.9

@AMIGASYSTEM

Quote:

AMIGASYSTEM wrote: @kolla I don't know, if you tell me I will thank you. I am not a developer

Same here, also not a developer, just a user.

The problem is that you try to write to the same file twice, and that doesn't work and you get an empty file as result. On "real" Amiga shell you even get error message. If you write to two different files instead, and try with different versions of lha, you can see what is stdout and what is stderr.

To move stderr to the stdout stream, you can on Amiga shell use ´*><´, so...

c:lha *>< >ram:output.txt

Last edited by kolla on 29-Apr-2024 at 02:57 AM.
Last edited by kolla on 29-Apr-2024 at 02:56 AM.
Last edited by kolla on 29-Apr-2024 at 02:54 AM.
Last edited by kolla on 29-Apr-2024 at 02:54 AM.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@Matt3k

Quote:

It certainly didn't help that Gateway was forced to dump Amiga by MS, that was probably the last real shot it had for some future regardless of what and who did what afterwards...

And also, though it seems to have escaped under the radar, Acer is also involved with the Amiga assets at some point.

@NutsAboutAmiga

Quote:

more precisely its emulated on top of x64 design. We should not be talking about x86, or Intel of the yesteryears, it’s not really relevant, “movbe� got me think about when did x64 designs became the norm.

Oh no. An x64 CPU emulating an x86 internally? The CPU world is turned on its head! Forget about discussions we see of FPGA CPU being simulation or emulation. Internal x86 CPU emulation will be where it's at.

@agami

Quote:

Yes. I would start with these as the low hanging fruit. Especially when one takes into consideration the logistics of the current OS4 development team.

PiStorm can do a fast 68K definitely. x64 certainly more so but not inside an Amiga yet. Fast x64 emulation of PPC is what's still lacking.

Quote:

After what I estimate to be better sales performance than that of the Hyperion Amiga OS 3 updates, someone might have that old idea anew and develop an Amithlon/Umilator style hypervisor layer for AmigaOS 4 on x64, or even ARM64.

The A600G looks to be proof of concept for this. Though its target is 68K it has ideas of replacing system components. Such as ARM optimised graphics. But, for fast PPC emulation analogous to Amithlon, a tight JIT emulator would be needed. I'm not aware of any PPC JIT but all I read about is Qemu and not much else. However I think perhaps a PJIT may be a better option. Or, my idea would be, to create a run time loader that treats PPC as virtual CPU like Java. So kind of like AmigaDE in some respects, but with an x64 loader, or ARM.

Quote:

And maybe one day the Apollo Team moves the 080 core to faster FPGA or realise their ultimate goals and tape out an 080 1GHz+ ASIC.

That was the plan. But will Facebook work on it by then? I've comments from Vampire fans they would like to do Facebook and other common web activities from their Vampires.

Quote:

How many AmigaOS 4 developers would there be after all or most that? At least a few more I'd venture.

A lot more if there was easier access to supported hardware.

@kolla

In that example, to set Path to the same location as Drawer but filtering out quotes.

The problem with your back tick solution is quotes will be saved to PATH but I suppose an Echo could be prepended before the back ticks to filter it out.

Quote:

Don’t create global variables in env: for stuff that like this, you end up with trouble when running same script many time in parallel, all trying to output to same file. Always use local variables in scripts. (And yell at Flype to fix RAChoice)

I prefer locals but had to figure why I did it that way. The same script shouldn't be run in parallel and it would result in windows opening up asking the user what to do. So why I did it that way is to preserve error output. As a subcommand the error output is lost. I don't know if the user pressed Ok and Cancel when what I get back is a blank string. Suppose specifying full path could work but I like to use a proper return code and not rely on a string.

@bhabbott

Quote:

Very few clones used 8086, and most of them weren't much faster because they still used the 8 bit expansion bus.

NEC V30 is an 8086 clone with 10% to 30% faster (depending on programs) due to internal improvements.

The IBM PS/2 models 25 and 30 were built with an 8 MHz 8086.

The Amstrad PC1512, PC1640, PC2086, PC3086 and PC5086 all used 8086 CPUs at 8 MHz.

The NEC PC-9801.

The Tandy 1000 SL-series and RL machines used 9.47 MHz 8086 CPUs.

The Olivetti M24/AT&T 6300 used 8086 8 Mhz to 8 Mhz.

The Compaq Deskpro used 8086 8 Mhz, sold as a faster XT clone.

The Wang PC used 8086 8 Mhz.

The Apricot PC used 8086 4.77 Mhz.

Sega's "16 bit" marketing wasn't the 1st.

Quote:

Not sure where you got this idea from. Compaq did not 'define' the 386. It used the same 16 bit bus as the PC-AT.

Compaq Deskpro 386 has a 32-bit front side bus and 32-bit memory despite 16-bit ISA slots. Refer to Amiga 2000 with the official C= 020 accelerator with 32-bit Fast RAM and 16-bit Zorro II slots for a similar solution.

Many 386DX PCs have 16-bit ISA slots. EISA was less popular on the desktop market i.e. Hint: ET4000AX SVGA ISA cards.

Quote:

On release there was no software that could use it in 386 enhanced mode.

The same can be said for AMD64. Windows XP X64 was released later i.e. April 25, 2005.

It's about the platform's road map and building the install base with the new hardware features.

Quote:

Microsoft Windows didn't go 32 bit until V3.0 in 1990, 4 years later.

As early September 1987, there's Windows/386 which introduced a protected mode kernel, above which the GUI and applications run as a virtual 8086 mode task. Windows/386 also provided EMS emulation, using the memory management features of the i386 to make RAM beyond 640k behave like the banked memory previously only supplied by add-in cards and used by popular DOS applications.

SCO released Xenix 386 for the 386 based PC in 1987. SCO's Xenix System V/386 was the first 32-bit operating system available on the market for the x86 CPU architecture.

386's built-in MMU sets the foundation for mass memory protected 32-bit OS development for the PC.

Quote:

IBM released the PS/2 Model 80 in April 1987, 7 months after the Compaq 386. It featured a 16-25MHz 386 CPU, VGA display, up to 4MB RAM onboard (expandable to 16MB with option board), 64k SRAM cache, 320MB SCSI hard drive, 1.44MB floppy drive, mouse port, and the 'new standard' 32 bit Micro Channel bus.

Micro Channel (MCA) bus is a failed standard. The rest of the PC market didn't care about MCA. ET3000 SVGA was released late 1987.

My classic Pentium Socket 7 has PCI and ISA slots.

My Dad dumped his IBM PS/2 Model 55SX due to the higher cost add-on cards. PS/2 Model 55SX had 16-bit MCA slots. MCA is not consistent with 32 bits.

The "gang of nine" made sure MCA fails.

EISA was also available on some non-IBM PC compatible machines such as the DEC AlphaServer, HP 9000 D-class, SGI Indigo2 and MIPS Magnum.

Last edited by Hammer on 29-Apr-2024 at 02:37 PM.
Last edited by Hammer on 29-Apr-2024 at 02:33 PM.
Last edited by Hammer on 29-Apr-2024 at 02:31 PM.
Last edited by Hammer on 29-Apr-2024 at 02:24 PM.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)

@matthey

Quote:

matthey wrote: OlafS25 Quote: Yes but user experience is in first line desktop. You can perfectly do something with the look & feel but sitting on top of something else. Also you could create layers that hide the different stuff (like linux) and make supporting or transition easier. A little like Apple did when they changed ISA to X86. I see no sense in the try to create another "modern platform" by reinventing all wheels. Besides we do not have the money for that. The Amiga is a successful chameleon. The Workbench desktop is not always what is seen first. Many games booted straight into the game and the CDTV and CD32 had special menus when not booting into other software. Some users have set up game systems that boot to a scrollable menu of games for easier access with a joystick/joypad. Some Amiga embedded software boots into a custom UI that may not be recognizable as an Amiga (kiosks for example). The Amiga 3000UX booted into Linux but a boot menu is available, as it is on most later Amigas, allowing boot into an AmigaOS partition. The Amiga has support for many input devices too including mouse, trackball, touch screens, light pens, drawing/sketch/art touch pads, joystick/joypad, paddles, etc. The Amiga can look like practically anything with input from practically anything. It can do this with tiny performance and memory by today's standards. It is easy to use and develop for. All this is valuable but it isn't used anywhere today because the hardware is too expensive. MorphOS adopted the Quark microkernel based on the L4 microkernel. QNX was nearly used to modernize or replace the AmigaOS. Gateway decided to replace much of the AmigaOS with Linux. Amiga Inc. tried to replace the AmigaOS with the AmigaDE (Amiga Digital Environment) virtual machine. ExecSG is a more organic development planned for AmigaOS 4. The AmigaOS needs some serious work under the hood to modernize it but what was extremely unpopular with Amiga fans was Gateway's decision to replace the AmigaOS with Linux and Amiga Inc.'s decision to replace the AmigaOS with AmigaDE. These radical choices solve the lack of modern drivers problem but a return to standard hardware greatly reduces the number of drivers required. I believe organic development and a return to standardized "custom" hardware would be a much more popular decision for Amiga fans than a replacement job or major parts under the hood. One of the major reasons why PPC AmigaNOne failed was because replacing the 68k with a PPC was like replacing an older but popular engine with a more modern but less popular one in a classic sports car. Classic sports cars with engine transplants are less valuable because they are seen as bastardizations. Rebuild the engine, replace internal parts with more modern higher tech parts, add a better flowing head and exhaust manifold, replace the old 8 bit EFI ECU and these tasteful mods are generally viewed favorably by fans. The AmigaOS is no different than the hardware. Most Amiga fans want an upgraded but compatible 68k Amiga and retro is hot right now. This may limit what is possible for the 68k Amiga but the 68k Amiga still has serious potential due to the choices for a dynamic and expandable PC by Jay Miner. Be careful not to throw out the 68k Amiga baby with the bathwater in pursuit of a modern Amiga.

AmigaPPC failed due to departures from A500/A1200's price range.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)

@Hammer

Good point. If you go back a bit further in time, the 500 and 2000 literally saved Commodore.

Again, Tom Rattigan made that call and brought them back from bankruptcy.

Having an entry and higher end offering is important.

The other issue is that in the 80's and early 90's there was still software being created by lots of parties for the Amiga (granted most of the business ones were poorly coded and had a short shelf life). Without some decent software and some backing making progress the appeal of the OS on expensive hardware will be less appealing to new users.

If the A1222 would have been priced like the Efica, it would have made sense that the X5000 was higher end and the A1222 be entry level, it would have helped.

To me the last issue is just the "tiredness" of it all, so many false hopes and failures. I still have fun but fun with actually doing work and hate tinkering like I used to. My PM G5 stay up for weeks at a time and just gets work done without any issues at a decent price point (Chris Edwards did a video a week or so ago with a similar system). It works and most of the programs I use get updated regularly, so I'm good and really don't care about "the" next hardware at this point...

Hypex Quote:

Oh no. An x64 CPU emulating an x86 internally? The CPU world is turned on its head! Forget about discussions we see of FPGA CPU being simulation or emulation. Internal x86 CPU emulation will be where it's at.

Sarcasm? Of course there is no x86 emulation or even simulation in a x86-64 CPU. Some of the x86 support may be microcoded but microcode is still hardware. If microcode support is emulation then the 68000 CPU was a big emulator which is just plain wrong. I thought about correcting Nuts too but Nuts will be Nuts.

Hammer Quote:

AmigaPPC failed due to departures from A500/A1200's price range.

PPC AmigaNOne failed because it was a major departure from the 68k Amiga in many ways. I believe the loss of 68k Amiga compatibility was a larger problem than early prices. For example, I believe a 68060 Boxer with PCI slots would have easily out sold early PPC AmigaNOne hardware and neither were or would have been cheap or offered value close to mass produced hardware. There are at least three 68060 accelerators currently available, at least one more planned despite a low supply of 68060s and AC68080 accelerators while there are no PPC accelerators available although the 68k Amiga and retro markets have expanded as also exemplified by THEA500 Mini success. Price and value are very important for hardware and likely also a major reason for the failure of PPC AmigaNOne hardware. It's not just that the AmigaNOne hardware is much too expensive and suffers from declining value due to old PPC silicon and bastard embedded CPUs, but it targets a different market for the elitist hobbyist classes which is practically the polar opposite of the 68k Amiga low price value market for the masses that C= targeted with the 68k Amiga. PPC AmigaNOne also lost the high tech value market as C= did by ignoring the high end market while cost reducing the 68k Amiga to turn it into the next C64. By the high tech value market, I mean the people who choose a product because they see it as state of the art tech that offers good value which the 68k Amiga represented in the mid to late 1980s. The PPC was also seen as high tech for awhile and early PPC AmigaNOne benefited from this but these types of people often move to newer products if the tech advantage and value are not maintained as was the case for the original 68k Amiga and the early PPC AmigaNOne hardware. It would be difficult and expensive to get the high tech value market back but the retro market based on nostalgia is available and hot right now. A simple cheap toy, even using emulation, satisfied this market but doesn't build a growing 68k Amiga user base and a sustainable Amiga market as we have seen by THEA500 Mini success. It is possible to over deliver expectations with hardware offering much more value using real hardware that could make the Amiga relevant again. Yes, it would cost millions of U.S. dollars to become competitive but maybe not that much more than the millions Trevor has already lost with the dead end PPC AmigaNOne.

@matthey

Good point there is very few uses for Classic Hardware with a PPC. You can play some games and draw the backdrop and images faster.

Most people sell their PPC since it does so very little.

AmigaOS4 wasn't that interesting to me on Classic and MorphOS is 50 releases behind on their last classic PPC offering.

To me, given the maturity of MorphOS it is a shame it wasn't still compatible, it would provide a bridge to faster hardware and give the old stuff something to do. 3.2x isn't any fun for me on Classic, as I still prefer 3.19 and there isn't any killer 68k programs to make it interesting either.

@Hypex

Quote:

Hypex wrote: @agami PiStorm can do a fast 68K definitely. x64 certainly more so but not inside an Amiga yet. Fast x64 emulation of PPC is what's still lacking.

The whole point of porting AmigaOS 4 to high performance 68k is to take PPC out of the Amiga equation.
If we have AmigaOS 4 on emu68 atop RPi 4 and RPi 5, and on UAE 68k atop x64, then it doesn’t matter how bad or almost usable the PPC emulation is on x64.

_________________
All the way, with 68k

@matthey

Quote:

matthey wrote: PPC AmigaNOne failed because it was a major departure from the 68k Amiga in many ways. I believe the loss of 68k Amiga compatibility was a larger problem than early prices. For example, I believe a 68060 Boxer with PCI slots would have easily out sold early PPC AmigaNOne hardware and neither were or would have been cheap or offered value close to mass produced hardware. There are at least three 68060 accelerators currently available, at least one more planned despite a low supply of 68060s and AC68080 accelerators while there are no PPC accelerators available although the 68k Amiga and retro markets have expanded as also exemplified by THEA500 Mini success. Price and value are very important for hardware and likely also a major reason for the failure of PPC AmigaNOne hardware. It's not just that the AmigaNOne hardware is much too expensive and suffers from declining value due to old PPC silicon and bastard embedded CPUs, but it targets a different market for the elitist hobbyist classes which is practically the polar opposite of the 68k Amiga low price value market for the masses that C= targeted with the 68k Amiga. PPC AmigaNOne also lost the high tech value market as C= did by ignoring the high end market while cost reducing the 68k Amiga to turn it into the next C64. By the high tech value market, I mean the people who choose a product because they see it as state of the art tech that offers good value which the 68k Amiga represented in the mid to late 1980s. The PPC was also seen as high tech for awhile and early PPC AmigaNOne benefited from this but these types of people often move to newer products if the tech advantage and value are not maintained as was the case for the original 68k Amiga and the early PPC AmigaNOne hardware. It would be difficult and expensive to get the high tech value market back but the retro market based on nostalgia is available and hot right now. A simple cheap toy, even using emulation, satisfied this market but doesn't build a growing 68k Amiga user base and a sustainable Amiga market as we have seen by THEA500 Mini success. It is possible to over deliver expectations with hardware offering much more value using real hardware that could make the Amiga relevant again. Yes, it would cost millions of U.S. dollars to become competitive but maybe not that much more than the millions Trevor has already lost with the dead end PPC AmigaNOne.

By 1996, 68060 wasn't competitive with the 1996 Pentium product stack.

PowerPC direction was done with German-located Amiga Technologies GmbH and Phase 5 GmbH. Amiga Technologies was stacked by ex-Commodore Germany's personnel. Commodore Germany wasn't the core logic design of Commodore.

Amiga's 68K-to-PowerPC approach mirrored Apple's move from 68K to PPC with a userland 68K emulator since Phase 5 is involved with PowerPC accelerators for the Mac market.

This Apple-influenced pathway wouldn't work for hit-the-metal Amiga games.

Apple established the color QuickDraw layer in 1987 as their RTG solution.

The Amiga 68K games are like PC's hit-the-metal i386 protected mode VGA games with 68K CPU and Amiga custom chips.

The direction should be Transmeta's method e.g. Emu68 pathway being the closest to it.

The Amiga is not a Mac.

Last edited by Hammer on 30-Apr-2024 at 07:17 AM.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)

@agami

Is emulation of PowerPC really so much worse than emulation of m68k?

Do you have side-by-side comparisons?

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC