@Raffaele

Quote:

Following your order of ideas we will never start any project!

Following your order of ideas some new weird project will start just in order to die very soon. First take a kernel which is well supported, then work on the rest. And trust me, linux is not that bad.

Quote:

First step is to migrate to new kernel even with its limited motherboard (or mobos)+cards+device already supported... Then we can build drivers for any kind of most used devices...

Who is "we"? The same "we" which needs years to support all hardware of X1000? The same "we" who very recently (finally!) provided stable USB2 support in OS4? The same "we" who still has lacks 3D or overlay?

Believe me or not, "building" drivers is not just ./configure && make...

@Raffaele

Quote:

Well, if monolithic kernels does not allow a device drivers structure just as like that we have in Amiga, why don't use hybrid kernels that are a viable existing solution?

Fine, fine one as well supported as linux ;)

Quote:

At least hybrid kernels don't require a complete recompiling of entire kernel to support every crap piece of hardware that user add to their computers...

WTF? I was using linux for years and had hardly to compile the kernel at all. Besides, on linux you can load the kernel modules on the fly anyway. No need to rebuild everything by every single hardware change. Actually I used also to change my *entire* hardware without any need to recompile the linux kernel at all.

Quote:

I remember I tried to compile Linux Kernel for my brand new PC motherboard that featured for the first time the USB ports, and USB support in Linux was only experimental and required kernel recompiling...

Come on, this is supposed to be an argument? Not so long time ago OS4 missed USB2 support, although it had the USB2 hardware and the standard was pretty old even that days. Does it mean OS4 sucks for you?

Let us agree Windows7 is crap because Windows95 had some serious issues many years ago. Let us agree OSX is crap because the old system for m68k was so limited. Let us agree OS4 is crap because AmigaOS1.2 was so crappy.

Quote:

Well... I never managed to make it working and I was forced to wait for stable new kernel releases almost 9 months.

Compare that to the first AmigaOne owners and their endless wait for OS4...

@Raffaele

Quote:

But why unfriendly monolithic Linux instead of the kernel you or me proposed?

Why not? And tell me what is so unfriendly in linux? Although it's monolithic it provides support for loadable modules. Moreover, it enables one to write userspace drivers, as I've pointed already.

Besides, if you like something else, join AROS Dev Team and write the damn code.

@michalsc

Quote:

Who is "we"? The same "we" which needs years to support all hardware of X1000? The same "we" who very recently (finally!) provided stable USB2 support in OS4? The same "we" who still has lacks 3D or overlay? Believe me or not, "building" drivers is not just ./configure && make...

We us amigans... And BTW I am using MorphOS not AmigaOS. But I am supporting all three sides.

It is not MOS team fault if AmigaOS lacked of USB 2 support for so long time...

We in MorphOS have it since 2.0/2.1.

But it is certainly fault of MorphOS Team for not having the courage of developing Quark Microkernel MorphOS to support Qbox together with ABox and provide a brand new Operating System with all modern features where Amiga Box should have been only a sandboxed enviroment...

Last edited by Raffaele on 01-Nov-2013 at 10:19 PM.
Last edited by Raffaele on 01-Nov-2013 at 10:16 PM.

_________________
"When the Amiga came out, everyone [at Apple] was scared as hell." (J.L. Gassée, former CEO of Apple France and chief of devs of Mac II-fx, interviewed by Amazing Computing, Nov 1996).

@Overflow

Quote:

Overflow wrote: @wawa Sometimes I do wonder what your agenda is. You seem to have a distinct dislike for PPC and Aos4 and generally Hyperions direction. While its not quite logical to invest in PPC/Aos4 if you compare it to the performance of Wintel machines, thats not really the crux of it. It would be like ridiculing Classical Amigans/C64 for spending money on upgrading their machines, tweaking and generally ...having fun! Same with Classical cars. Performance not up to par with current technologies, but people still enjoy it. Trevor/Aeon, Amigakit and Hyperion seems to enjoy their "universe" and quite a few people share that view. Let them enjoy it.

To me it is stranger that people are still supporting the choice of sticking with ppc. It has been pretty much a dead platform for quite some time that is so overpriced that only the most loyal amigans/fanboys can defend paying so much money for a underperforming system.

There just isnt any hope for any growth at all in userbase when you stick with ppc, which is something that is bad for you as well. Dont you want at least a chance for growth in the future and more software/games?? This just wont happen with ppc instead we are losing users for each day, which is something that can be seen just by looking at the activity of amigaworld these days.

Going to x86 not only gives us cheaper hardware, but it makes porting of software/games easier as well. With x86 it might even be possible to get something like wine or at least some linux compatibility like someone is trying to do with some AROS spinoff.

Now the situation is so bad that there are currently no web browser even in development that support modern standards. The outdated OWB-mui is a complete joke compared to the more updated MOS version and timberwolf even more so..

@Raffaele

Quote:

But why unfriendly monolithic Linux instead of the kernel you or me proposed?

Linux has enormous advantages in the areas of drivers and application software — the most recent versions of LibreOffice, Firefox, Chromium, Blender, etc. And Valve games!

There's nothing wrong with microkernels, especially if you're a systems programmer and enjoy working at that level. But I would get a decent Amiga-like userland up and running on Linux or BSD first so that users have some decent application software, and then go back and work on the kernel.

In other words, build the system top-down, starting with the UI and apps, and work down from there.

For what it's worth, I've been using Linux for 15 years, and it's been pretty good. I've had one kernel panic in all that time. I like if better than OS X, which I use at work, and I like it a lot better than Windows, which I use as little as possible.

Last edited by bison on 02-Nov-2013 at 03:28 AM.

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

So... just use Linux!

There are other o.s.es which have more drivers, if this is the primary reason.

Aside that, I fail to see the added value of such Frankenstein. Especially comparated to the already existing AROS hosted versions.

@bison

This discussion started how to choose an alien kernel to AmigaOS environment and how to make Amiga experience of ease of use and user friendly being the same, and it ends like usual with people aiming at new software always pointing at thingies like Firefox that is not the best browser ever... but it is only the most known and buzzed up in geek gossip sites...

Listen to me... I do not care a bulls**** that Linux kernel it is not to bad and it supports plenty of hardware and its OS of reference Ux/Linux has dozzilions of software...

The problem is not adapting Linux kernel that offer support for a wide range of hardware solutions...

The problem is really to found a friendly kernel, and sure hybrid kernels such as Be/Haiku are more similar to Amiga approach to computing...

Then it is necessary to tie this kernel with existing park of hardware: AmigaOnes, Samanthas, Pegasoses and Amiga X1000s hardware and make it running well supporting in a sandbox all existing Amiga good software not to loose users without even basic computing capabilities.

Look at MacOS X... it is intimely tied with the limited hardware provided by Apple... And Apple make a great effort to make kernel initializing and supporting only that hardware in order to prevent driver faults, make hardware running at its best, and make life easy even for inexperienced Apple MacOs users.

That's the way we must follow: Choose the most Amigaish kernel (not the most bloated Linux one), and make it talking intimely well with the rest of the OS (device handlers structures, consoles, libraries, DOS, GUI interface, Filesystem).
In the beginning it must support... (and IT HAS TO) only a limited number of motherboards and devices in order to make it working flawlessly.

Once we made this, thanks to new kernel we could add all the things that lacks in Amiga: i.e. protected memory, MMU management, multiuser capabilities and SMP.

In the future then we could migrate to different processor architectures and adding support for more hardware devices.

-----

All you readers wash your ears and your eyes carefully and remember well this point:

Adding a new kernel does make Amiga POSIX compliant, that's it...

BUT DOES NOT MAKE IT AUTOMATICALLY COMPATIBLE WITH OTHER OPERATING SYSTEMS LIKE UNIX/LINUX.

Sure AmigaOS powered with a Unix/Linux/BSD kernel will make it easier to port POSIX compliant software, but then the porting difficulties remain more or less the same because you must ADAPT this software to Amiga GUI STYLEGUIDES and TEACH IT how to deal with AMIGA FILESYSTEMS.

Last edited by Raffaele on 02-Nov-2013 at 11:23 AM.
Last edited by Raffaele on 02-Nov-2013 at 11:22 AM.
Last edited by Raffaele on 02-Nov-2013 at 11:18 AM.
Last edited by Raffaele on 02-Nov-2013 at 11:06 AM.
Last edited by Raffaele on 02-Nov-2013 at 10:54 AM.
Last edited by Raffaele on 02-Nov-2013 at 10:49 AM.
Last edited by Raffaele on 02-Nov-2013 at 10:48 AM.
Last edited by Raffaele on 02-Nov-2013 at 10:42 AM.
Last edited by Raffaele on 02-Nov-2013 at 10:31 AM.
Last edited by Raffaele on 02-Nov-2013 at 10:28 AM.
Last edited by Raffaele on 02-Nov-2013 at 10:27 AM.
Last edited by Raffaele on 02-Nov-2013 at 10:24 AM.
Last edited by Raffaele on 02-Nov-2013 at 10:23 AM.
Last edited by Raffaele on 02-Nov-2013 at 10:21 AM.
Last edited by Raffaele on 02-Nov-2013 at 10:18 AM.
Last edited by Raffaele on 02-Nov-2013 at 10:17 AM.

_________________
"When the Amiga came out, everyone [at Apple] was scared as hell." (J.L. Gassée, former CEO of Apple France and chief of devs of Mac II-fx, interviewed by Amazing Computing, Nov 1996).

It doesn't make sense, and for several reasons. The primary one is that the post-Amiga o.s.es are simply a matter of HOBBY. Yes, they are developed for fun (even by the developers themselves), to make a bunch of people remind the good old Amiga days. Even if they are sold as commercial products, they still remain hobby products.

To be clear, you don't have to compete with colossus such as Microsoft and Apple. Not even dwarfs like Linux, which in any case are immensely greater than the post-Amiga o.s.es as audience.

So, IMHO it doesn't make sense to throw away much of what makes unique an Amiga o.s./like, just to "fill the gap", because users are frustrated by the chronic lack of modern o.s.es features, such as memory protection, resource tracking, SMP, and drivers ultimately.

Instead of taking the opportunity to experiment and propose new, "creative", solutions to current issues, you just want to follow the easy way and uniform the existing to what is already found outside. But if you look outside, there are HUNDREDS of o.s.es which are Unix/Posix compliant.

It seams that if your o.s. is not Unix/Posix-compliant nowadays you're not in the "right" way. Research? Nah! It's something academic and, BTW, it's still well Unix/Posix-centered. There's no space for new, stimulating, innovative ideas. Everything should be flattened to the Unix/Posix positions, because, you know, it's the holy Grail of the o.s.es....

The Amiga o.s. was new? Innovative? Proposed new ideas? Who knows. Or... who matters, now? Go ahead with the "right" way that others have already traced: it's the "right" one. For sure!

In the end, the easy way is to get a kernel, throw away the old one, and put something on top of it. Oh, yeah, it's something which nobody has already seen. An incredible, brilliant, outstanding idea...

And to get the old software still work on the "new" environment, you think to put it running inside a (well sealed) sandbox. What a genius! An unbelievable stuff!!! It's called innovation, yeah! Don't you feel the good old Amiga days?

Now turn on your PC (or Mac, which IS a PC), and run (Win)UAE. You've a modern (sic!), solid, SMP-capable, and with plenty of drivers, o.s. which runs the old Amiga software in a sandbox. What a miracle!!!

But what's missing? The Amiga look and feel, primarily. And is there any solution for this? No, using a theme is the easy way, which is not suitable for geeks/nerds. The not-so-hard way is to just replace the so called Desktop Environment with a new Amiga-like, after the kernel booted. Something which other coders have already made (a quick search will prove it). So it's perfectly feasible. Now. And without a great effort. If that's what you want.
And for coders, wrapping the old APIs into some library (or set of libraries) doesn't require that much effort. Again, if that's what you want.

But, please, stop proposing a Frankenstein putting together two things that are completely different. It's enough having the post-Amiga o.s.es Unixizating them due to the introduction of (partially) shared objects (libraries? Too hard!), interfaces (a single interface to rule them all? Good only for a novel!), lib* (datatypes? WTF!), X11 (a socket to communicate with the graphic stuff. Stunning!), Qt (Intuition, gadtools, BOOPSI? Too hard!), and other software which is already available (and running well)...

@cdimauro

Quote:

So... just use Linux!

Do you mean "Use the Linux kernel as the basis for a new Amiga-like OS," or "Use an existing Linux distribution?"

The former is what I am proposing. Or actually, others have proposed, and I agree with as being a sound plan.

The later is what I am already doing, but it is not ideal. The system takes too long to boot, and there is too much latency in the user interface. I'm also not a big fan of bash, or any other Unix shell (with the exception of rc from Plan 9). These things can be fixed, which is what I'm advocating. No current Linux-based system is very Amiga-like, although Tiny Core comes impressively close in some respects.

Quote:

There are other o.s.es which have more drivers, if this is the primary reason.

There are other systems, such as Windows, that have more drivers, but they are not free/open source, so you can't use them as the base for a new OS. Linux currently has the best driver support of any of the available free/open source operating systems.

Quote:

Aside that, I fail to see the added value of such Frankenstein. Especially comparated to the already existing AROS hosted versions.

I run Aros (and FS-UAE) on Linux, and I like it, but it's not my everyday system because of the lack of application software. And there's very little incentive for people to write application software for Aros, so the situation will probably not improve much.

Application software is a big lever. If you can create a system that other people want to create software for, that will greatly increase its chances of success.

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

@Raffaele

Quote:

This discussion started how to choose an alien kernel to AmigaOS environment and how to make Amiga experience of ease of use and user friendly being the same, and it ends like usual with people aiming at new software always pointing at thingies like Firefox that is not the best browser ever... but it is only the most known and buzzed up in geek gossip sites...

Agreed, Chromium is arguably a better browser than Firefox. Maybe some others as well. But Firefox is still a propular browser, and many people use it.

Quote:

Listen to me... I do not care a bulls**** that Linux kernel it is not to bad and it supports plenty of hardware and its OS of reference Ux/Linux has dozzilions of software...

Application software support is important. Without it, a system has almost no chance of success.

Quote:

The problem is not adapting Linux kernel that offer support for a wide range of hardware solutions... The problem is really to found a friendly kernel, and sure hybrid kernels such as Be/Haiku are more similar to Amiga approach to computing... Then it is necessary to tie this kernel with existing park of hardware: AmigaOnes, Samanthas, Pegasoses and Amiga X1000s hardware and make it running well supporting in a sandbox all existing Amiga good software not to loose users without even basic computing capabilities.

Exisiting hardware is one of the things holding AmigaOS back — it's all scarce, expensive, and underperforming compared to commodity hardware.

Quote:

Look at MacOS X... it is intimely tied with the limited hardware provided by Apple... And Apple make a great effort to make kernel initializing and supporting only that hardware in order to prevent driver faults, make hardware running at its best, and make life easy even for inexperienced Apple MacOs users.

Apple OS X is a great example of building a new system on an existing base system. Most OS X users have no idea that they're running Unix, and even if they do they don't know the significance of it. The base system doesn't really matter that much, so long as it's fast and reliable.

Quote:

That's the way we must follow: Choose the most Amigaish kernel (not the most bloated Linux one), and make it talking intimely well with the rest of the OS (device handlers structures, consoles, libraries, DOS, GUI interface, Filesystem).

"Linux is bloated" is one of the great urban legends of our time. The source tree is "bloated" — it's huge, in fact — and the kernels that ship with most distributions are large because they want to support as much hardware "out of the box" as possible. But much smaller kernels can be compiled to support very limited systems as well.

Quote:

In the beginning it must support... (and IT HAS TO) only a limited number of motherboards and devices in order to make it working flawlessly.

Yes, I agree; this is a good idea. If Linux is used, it should be compiled to support a reference system, and be as small as possible.

Quote:

Once we made this, thanks to new kernel we could add all the things that lacks in Amiga: i.e. protected memory, MMU management, multiuser capabilities and SMP.

You can't really add those things without refactoring a huge amount of code and breaking APIs. They have to be part of the initial design.

Quote:

In the future then we could migrate to different processor architectures and adding support for more hardware devices.

Yes, more hardware devices for sure. As far as hardware architecture, x86 and ARM are probably the only two that matter much anymore.

Last edited by bison on 02-Nov-2013 at 06:39 PM.

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

@thread

Once when QNX kernell was to be used... it was very interesting to read how you do not ever have to reboot QNX. You can even update the kernell "on the fly".

So, surely Linux kernell would be interesting because of the drivers availability, but is it the world's most important thing?

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

@KimmoK

Quote:

KimmoK wrote: So, surely Linux kernell would be interesting because of the drivers availability, but is it the world's most important thing?

No.

Sounds like being 'Stuck in a rut' .

_________________
Tinkering with computers.

@KimmoK

Quote:

KimmoK wrote: So, surely Linux kernell would be interesting because of the drivers availability, but is it the world's most important thing?

_________________
AmigaOS or MorphOS on x86 would sell orders of magnitude more than the current,
hardware-intensive solutions. And they'd go faster.-- D.Haynie

@Hondo

Simple answer no!

_________________

@bison

Quote:

Agreed, Chromium is arguably a better browser than Firefox. Maybe some others as well. But Firefox is still a propular browser, and many people use it.

Maybe if you like less features and more memory useage when having alot of tabs open.
I dont really get why people still dislike firefox, considering that they have fixed all the memory leaks of the past.
I have tried using both chrome and chromium and to me it dosent even come close to firefox.

@Tomas

Quote:

Tomas wrote: @bison [quote] I dont really get why people still dislike firefox, considering that they have fixed all the memory leaks of the past.

>For example Firefox for Windows: I don't like the way it deals with Tabs, downloads and bookmarks.

An example: Any time you had to add a bookmark it presents a minimal menu... Very ugly!

Also it makes not easy to understand if you created bookmark into the new drawer you want or not. You had to enter then bookmarks menu and check if you placed it in the right place.

Opera is more friendly either with tabs, downloads and bookmarks. Unfortunately when they abandoned their engine "Presto" for Webkit, they made an ugly bugged mess and Opera now it is almost unusable.

BUT WAIT A MOMENT:

Opera for Android, despite the fact it is ugly bugged has its own text input window that is far better the best text editing window unsurpassed by any other Android Browser.

Why? Because other browsers like the Android standard one, Dolphin Bowser or Firefox, use the system input window that is buggy, and makes had to input text more than 5 lines, as cursor is uncapable to going up and down in the right place.

While Firefox is the only browser on any platform, capable to deal with enormous Facebook news pages, when you continue digging news and news...

Well I must admit that when on Windows, I use Firefox just only for facebook and Firefox is good only to browse facebook enormous and long news page...

_________________
"When the Amiga came out, everyone [at Apple] was scared as hell." (J.L. Gassée, former CEO of Apple France and chief of devs of Mac II-fx, interviewed by Amazing Computing, Nov 1996).

And again, the reason you keep bothering about drivers, kernels and other bits of stuff which is far from the User Experience, is because for some reason you can't think anymore of the Amiga as it was: a vertically integrated solution.


A recipe for not wasting 10 more years:

- Build, say, 4 X86 based computers based on different configs ranging from cheap to high end
- Fit them into some well designed case which sets itself apart from the other stuff out there
- Take a solid kernel. Say, Linux. Strip it out of everything that is not relevant to the hardware inside just those 4 configs. Add UAE.
- Now work on the desktop/ui/blabla using all those Amiga signature elements we all keep mentioning again and again and again

If this "recipe" reminds you of something, it's because it's exactly what Apple did to bring back their desktop stuff from oblivion.

It's something that could be executed with some coordination, even by just a team of 4 or 5, with the case design and the UI/UX work being the parts where some investment would be required.


But it's too simple, isn't it? No one believes in "simple" anymore - and that's how we landed in the age of measuring a CPU's temperature, choosing between 30 different graphic card looking at which one scores 1 more frame or 1 less db of noise... and arguing about kernel designs.

@Tomas

Quote:

I dont really get why people still dislike firefox, considering that they have fixed all the memory leaks of the past.

The primary thing I don't like about Firefox is how a runaway script in one tab can freeze the entire UI for minutes, or even indefinitely. Sometimes it's so bad that it won't even redraw its own window. I've had this happen dozens of times on both Linux and OS X, and just two days ago on Firefox 25, so it's an unresolved problem — probably a architectual problem that is very hard to fix.

Chromium probably has issues of its own, but I don't use it enought to know what they are.

Last edited by bison on 03-Nov-2013 at 03:31 AM.

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

Quote:

bison wrote: @cdimauro Quote: So... just use Linux! Do you mean "Use the Linux kernel as the basis for a new Amiga-like OS," or "Use an existing Linux distribution?" The former is what I am proposing. Or actually, others have proposed, and I agree with as being a sound plan.

And which I don't like.

Quote:

The later is what I am already doing, but it is not ideal. The system takes too long to boot, and there is too much latency in the user interface.

I suspend my machine, both because I have no boot time and I find back the environment exactly as I left it, which has a BIG value (at least for me). I reboot only once per month, if I get new updates that require it; or because I want to use Icaros natively. Except for these cases I never reboot, so the boot time doesn't belong to or affect me.

Quote:

I'm also not a big fan of bash, or any other Unix shell (with the exception of rc from Plan 9). These things can be fixed, which is what I'm advocating.

The same is for me. I like the Amiga shell. Or, alternatively, I greatly prefer a Python interactive shell.

Quote:

No current Linux-based system is very Amiga-like, although Tiny Core comes impressively close in some respects.

I don't know Tiny Core, but I've a general Linux knowledge, and I don't like it.

Quote:

Quote: There are other o.s.es which have more drivers, if this is the primary reason. There are other systems, such as Windows, that have more drivers, but they are not free/open source, so you can't use them as the base for a new OS. Linux currently has the best driver support of any of the available free/open source operating systems.

http://reactos.org/it
It's free. It has the same Windows API & ABI. So it can use the existing Windows drivers base, which is pretty HUGE and of good quality.

Quote:

Quote: Aside that, I fail to see the added value of such Frankenstein. Especially comparated to the already existing AROS hosted versions. I run Aros (and FS-UAE) on Linux, and I like it, but it's not my everyday system because of the lack of application software. And there's very little incentive for people to write application software for Aros, so the situation will probably not improve much. Application software is a big lever. If you can create a system that other people want to create software for, that will greatly increase its chances of success.

Which is very difficult nowadays.

Anyway, I prefer to wait and see what comes out from ARIX. At least for my curiosity.