@BigD

Quote:

Whose going to make the decision to scrap Exec? Thomas Richter, Trevor, Matthew from AmigaKit?

It may have escaped your attention that Trevor Dickinson has recently acquired ExecSG, and a team has been formed to work on the kernel. So the chance that Trevor or AmigaKit will scrap Exec some time soon is rather slim.

_________________
The Rear Window blog

AmigaOne X5000/020 @ 2GHz / 4GB RAM / Radeon RX 560 / ESI Juli@ / AmigaOS 4.1 Final Edition
SAM440ep-flex @ 667MHz / 1GB RAM / Radeon 9250 / AmigaOS 4.1 Final Edition

@Trixie

Quote:

It may have escaped your attention that Trevor Dickinson has recently acquired ExecSG, and a team has been formed to work on the kernel. So the chance that Trevor or AmigaKit will scrap Exec some time soon is rather slim.

Bug what use is a Second Generation Exec when you want Third Generation Amiga?

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla

You increase the numeral by one

_________________
The Rear Window blog

AmigaOne X5000/020 @ 2GHz / 4GB RAM / Radeon RX 560 / ESI Juli@ / AmigaOS 4.1 Final Edition
SAM440ep-flex @ 667MHz / 1GB RAM / Radeon 9250 / AmigaOS 4.1 Final Edition

@thread

Is there nowdays possible to build native drivers for "AROS on ARM" -kind of solution?
Is RPi4 GPU open to develop drivers for (without billion$ for dev docs).

IMO, Raspberry Pi 4 Model B would seem to have capability to deliver compareable/better experience vs my SAM440 (&vs Tabor & vs MorphMiniG4).

((I'm not sure if Amiga SW will need A64/x64 kind of power ever. We just need fun and useful system to "tinker" with.))

_________________
- KimmoK
// For freedom, for honor, for AMIGA
//
// Thing that I should find more time for: CC64 - 64bit Community Computer?

...

>Should Amiga-like systems abandon the Exec.library kernel?

exec compatibility is not needed, if it limits future development.
Sandbox can be made for legacy, again.

>Should Amiga-like systems go bi-endian like AROS? Little endian?

If we can afford, yes, at least as long as most current PPC machines are alive, then -> little endian.
If we can not afford -> LittleEndian, today, please.

>Should we try to maintain cross-compatibility via some bytecode so we don't need to care about the OS at all?

AmigaOE/AmigaDE/intent again .... ?
Scary, but OS + driver professionals can say if it is a sane way.

((...WinUAE is not too bad... not to mention berniethlon ... so not everything need to bang the HW directly ...))

>Who will supply drivers?

Good question.
Should we be able to boot without PCIe GFX+Audio?
Can we have native AmigaOS5 RPi4 drivers? How?
We have RadeonHD competence, how easily it can be used for third generation?

Last edited by KimmoK on 01-Oct-2019 at 12:33 PM.

_________________
- KimmoK
// For freedom, for honor, for AMIGA
//
// Thing that I should find more time for: CC64 - 64bit Community Computer?

Quote:

Samurai_Crow wrote: As for the touchscreen interface, it would be a driver issue if you tried for a desktop version.

You don't get it, Touchscreen GUIs and Mouse GUIs are totally different.

If you don't believe, just try to use a touchpad of a laptop without using the mouse buttons it is no fun, your mouse is suddenly reduced to a one button mouse where the mouse button is always pressed during mouse usage. Most applications will need a redesign, especially when buttons where designed for the precise usage of a mouse (or a pen) instead a broad and bulky finger.

That's why Apple had success with iOS on tablets where Microsoft failed with Win3.11, Win95 or WinXP on tablets. It was not a 'driver thing' it was a GUI and application thing.

And for Android, it has a Linux Kernel but is using a totally different GUI than a desktop Linux.
See: https://en.wikipedia.org/wiki/Android_(operating_system)

Some people had fantasized about AmigaOS on a Mobile, well the AmigaOS is small but rather outdated when it come to security and support of multiples cores but its GUI is not usable for a mobile phone or a tablet, changing the GUI would lead to something with a look and feel of Android or iOS. Amiga users would not recognize it as Amiga system.

Last edited by OneTimer1 on 01-Oct-2019 at 03:08 PM.

@OneTimer1

Quote:

You don't get it, Touchscreen GUIs and Mouse GUIs are totally different.

Very true. My laptop has a touchscreen, and I hate it. But am sortof lucky to have it as sometimes a Windows 10 update kills my mouse completely, and the touchscreen allowed me to recover that, as awkward as that is.

I have a tendency to point at the screen to keep my place for certain tasks, and the touchscreen then wants to go and do some responsive sort of thing, when I do not want it to. Annoying and confusing, and makes me wish I had not got the touchscreen.
It is also very glossy and glarey, I wish I had a matte antiglare screen.

This morning, my son changed it to tablet mode, and I could not even use the laptop until he said what happened and where the setting was for me to change it back to normal laptop mode.


I have a Dell Precision laptop, and they keep trying to install a 3rd party mouse driver from Alps, which is aweful. It causes significant lag, on the order of seconds, making it terrible to click and drag anything. As I have 10 Pro, I could create a policy to prevent installing the Alps driver, which is very preferrable. The problem is that when there is a major 10 feature update, they delete the default MS mouse driver, which works great for me, and then I have no mouse driver at all. Back in Windows 7, the same thing allowed me to prevent Alps driver from loading, but it did not remove the default driver, so it remained useful until it updated me to 10, and now I have to toggle that policy on and off after major updates, and that is not a trivial thing to do, but at least is possible with the touchscreen working... I don't know if I would be able to recover from only a keyboard.

_________________
All glory to the Hypnotoad!

@Trixie

Quote:

It may have escaped your attention that Trevor Dickinson has recently acquired ExecSG, and a team has been formed to work on the kernel. So the chance that Trevor or AmigaKit will scrap Exec some time soon is rather slim.

Has that been confirmed? I've seen the rumors for a while, and Trevor saying that Aeon does not have it, but have not seen any confirmation that Revor personally, or some other of his organizations bought it.

_________________
All glory to the Hypnotoad!

@kolla

Quote:

Bug what use is a Second Generation Exec when you want Third Generation Amiga?

ExecSG looks the same as ExecSG.

@outlawal2

Quote:

outlawal2 wrote: @megol Regardless, if you had to do anything under the hood (and you often do especially with Amiga) then you would still be hit in the face with Linux.. And again I am not interested in Linux unless I want to do something with Linux

What exactly do you fear when doing anything low level? Yes you would have to design drivers using Linux interfaces and a few parts of the root directory may be hardcoded (don't really know). But even then the root could be hidden from users just as it is in Windows NT where the actual root of the system is \\ and all drive letters are mapped onto this in user level.
You wouldn't need to use an Unix shell or Linux specific commands to do things - it could all be handled in user level code. The result could look and feel like a modern Amiga OS.

@simplex

From the "Getting outside programmers interested" thread:

Quote:

In the decades I've been following this argument on forums, I've found that people disagree on the philosophy; an apparently vanishing number think the custom chips are the essence of an Amiga (15 years ago this group seemed dominant); others, the ability to bang the metal; still others, the peculiarities of the OS (which today are generally regarded as deficiencies, though in the 80s they were acceptable compromises). This divergence of opinion strikes me as a logical consequence of the machine's original conception, precarious gestation, and eventual birth. If one accepts my view of what makes the Amiga philosophy, then there's an argument to be made that every "personal" computer today is an Amiga: preemptive multitasking, custom chips for video, libraries to provide APIs, more than one way to interact with the system (both graphical and command-line), etc. I'd much rather live in that world than the aut/aut world we had when the Amiga came out: monochrome "serious" machine or colorful game machine, graphical or command line, single-tasking home computer or multitasking workstation, etc. But in some sense, that philosophy predates the Amiga; it's just that Amiga was the first to bring the package to market, and unfortunate enough to be bought by Commodore instead of someone who would see it as their core business.

My abstract definition (ignoring all hardware and software details) of the "Amiga philosophy": Allow the user to be in complete control of their computer and use it however they want for whatever they want. This necessitates that the system be as simple as possible, and as open and extensible as possible. The capabilities-to-complexity ratio of the Amiga when it first came out was much higher than what is generally available today.

_________________
"Unix is supposed to fix that." -- Jay Miner

@megol

Quote:

But even then the root could be hidden from users just as it is in Windows NT where the actual root of the system is \\ and all drive letters are mapped onto this in user level. You wouldn't need to use an Unix shell or Linux specific commands to do things - it could all be handled in user level code. The result could look and feel like a modern Amiga OS.

Let's wait till Microsoft are done abstracting Windows away from NT kernel and instead offer it as a user environment "service" that runs on several different underlying kernels, including Linux, and then the "Amiga peeps" can go "oooh... we should do that!"

BTW - PowerShell should be ported to AROS.

Last edited by kolla on 02-Oct-2019 at 10:17 AM.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla

Quote:

Let's wait till Microsoft are done abstracting Windows away from NT kernel and instead offer it as a user environment "service" that runs on several different underlying kernels, including Linux, and then the "Amiga peeps" can go "oooh... we should do that!"

Yeah.

There are more than a few among us who can't seem to discern the difference between Linux, Xorg, Gnome, bash, vim, and who knows what else.

Last edited by bison on 02-Oct-2019 at 05:17 PM.

_________________
"Unix is supposed to fix that." -- Jay Miner

@bison

Well it more like it goes in other direction, its moved from the OS to the WEB browser, so yep you can run some Microsoft stuff there, but it's generally slower, and web version often has less features.

Last edited by NutsAboutAmiga on 02-Oct-2019 at 05:22 PM.
Last edited by NutsAboutAmiga on 02-Oct-2019 at 05:21 PM.
Last edited by NutsAboutAmiga on 02-Oct-2019 at 05:20 PM.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

Re:Frameworks for development

While object-oriented frameworks like .NET are common, the introduction of overhead from the ever-expanding class library should not be ignored. C# and PowerShell are elegant as long as the framework is limited to abstract datatypes. Once OOP starts creeping into one-time use code....

As an example, I'm making an adventure game library in AmigaE using OOP and found that the function pointer calls needed to do interface inheritance were never fully debugged. How did that happen? E isn't often used for frameworks of any great depth. Single inheritance can be overkill already for small projects like Amigans tend to do. I'm probably the first person to even need an interface class because it's beyond the scope of most Amigans programming habits to write big frameworks.

If we want to get Amiga coding to where modern development is, we'll need to get back to the basics first. Code reuse can improve cache performance if abstraction overhead isn't bad.

@kolla
Microsoft can do that and Amiga can move atop of the NT kernel instead. Most crap in the system comes from the Win32 subsystem and the kernel is a pretty good design. We'd get driver support for almost every type of hardware out there too.

@NutsAboutAmiga

Quote:

Well it more like it goes in other direction, its moved from the OS to the WEB browser, so yep you can run some Microsoft stuff there, but it's generally slower, and web version often has less features.

I assume this is in response to post #53?

What I mean is this: Someone (we'll call him Fred) tries a Linux distro for the first time and has trouble with bash (as one does), so he says "I hate Linux." But what Fred really hates is bash, not realizing that bash is not an integral part of Linux. The same goes with Xorg, Gnome, etc. A lot of what people dislike about Linux isn't even Linux -- its all the stuff added on top to make a complete, functioning system.

I personally like the idea of using Linux, with its drivers, filesystems, network stack, etc., and replacing the rest with more Amiga-like components. This is achievable, since all the hard stuff if already done.

_________________
"Unix is supposed to fix that." -- Jay Miner

Re:AROS Hosted

Let me pose the question: Would Hosted AROS be sufficiently backward compatible as an emulation layer for older software? It's a superset of 3.1 with irrelevent, obsolete features removed. It's the only second generation Amiga-like that's backwardly compatible to 68k at all in the ABI v1 edition.

Re:AmigaVM

If we rebuilt AmigaDE as a superset of WebAssembly we'd need a 32-bit and 64-bit version and bindings for big and little endian architcture for each. I like this idea mainly because the performance will leave JIT compilation in the dust as much as WebAssembly is likely to leave JavaScript in the dust as well. Emscripten is already written and recompiles its own source into WebAssembly also.

Re:future vision

Continuing the line of products that once was known as Amiga, I can only see one affordable path: cluster computing. There are "compute unit" cards that put the capabilities of a RasPi3 on a $20 USD card. It is self contained enough to include boot media and 1 GiB of RAM on each. Now imagine running 8 in parallel. That's 32 cores and 8 GiB of RAM for $160 plus the motherboard they are plugged into. If that board were plugged into an add-on card for the PineBook Pro laptop that would be 38 cores and 12 Gigs of RAM for $400.

The downside is that the units need to talk to each other over USB2 lines to have enough connections to propogate as a hypercube network and a new version of Rust or Go that can cope with the parallelism of 8 slave busses and one master bus.

Is there enough interest or energy to do this?

@Samurai_Crow

Why would there be the need for such hardware when we still don't have SMT? What 'Amiga' software would be scalable to use all those cores! The Cell was a tough sell (no pun intended ) despite being great at off loading tasks to the SPEs!

The investment in powerful hardware (even if relatively cheap if not factoring software support costs / drivers etc) must be preceeded by a period of stablisation, user growth and actual software porting / production IMHO! That's what the Tabor was for! It's all the more painful that it's limping out of the door!

Last edited by BigD on 04-Oct-2019 at 08:31 AM.

_________________
"Art challenges technology. Technology inspires the art."
John Lasseter, Co-Founder of Pixar Animation Studios