@falemagn

Just a question about AROS in hosted mode, can AROS access hardware that is not supported in the hosted OS ? I.e. if I don't have the Linux USB kernel module, can I still access USB peripherals from AROS (provided that there is AROS driver for this peripheral) ?

_________________
AOS 4.1 : I dream it, Hyperion did it !
Now dreaming AOS 4.2...
Thank you to all devs involved for this great job !

jahc said: Quote:

My AmigaOne is the super Amiga I could never afford when I was growing up. I prefer to use this machine over a PC because I'm in love with the OS, and enjoy doing the every day stuff (like email) on AmigaOS. The feel of the platform is awesome. We just need some new hardware and a bit more software. As it is though, the platform covers most of my needs already.

I can completely use those words too !

_________________
Rock lobster bit me. My Workbench has always preferences. X1000 + AmigaOS4.1 FE
"Anyone can build a fast CPU. The trick is to build a fast system." -Seymour Cray

@abalaban

Yes, it can. As a matter of facts, the PCI subsystem works under linux, provided you have root priviledges, as well as on the bare hardware.

_________________
“It is well enough that people of the nation do not understand our banking and monetary
system, for if they did, I believe there would be a revolution before tomorrow morning.”
~~ Henry Ford

@falemagn

so what the advantage of being hosted for AROS ? Except the fact that it can't crash the whole machine (beeing launched as a task of the underlying OS) and that it does not have to boot the machine.

_________________
AOS 4.1 : I dream it, Hyperion did it !
Now dreaming AOS 4.2...
Thank you to all devs involved for this great job !

@falemagn

Quote:

@abalaban Yes, it can. As a matter of facts, the PCI subsystem works under linux, provided you have root priviledges, as well as on the bare hardware.

I don’t know how useful that is under AmigaOS, when I wrote the cw.device, it where just matter of soft rebooting and the new device drivers where loaded automatically.

I don’t think AROS will success if the main aim is live in side Linux, AROS need the applications native, or else it just becomes Linux distribution whit an strange desktop on top, anyway Linux boots slowly.

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

@falemagn

Quote:

So, there's nothing inherent to MacOSX that would make you prefer it over BSD? Is it just a matter of how it's commercialized and whether or not there's some kind of standards, whatever it is?

Not really, no. Maybe it's because I'm not a member of the Mac community and see Macs as just computers, but as I said in another post, Macs are OK but nothing special. Maybe it's because I've used Linux for years, and so I can see through Apple's marketing to see what MacOS X really is: a set of libraries and a neat window manager on top of a Unix clone. Cool APIs to do just about anything and neat window managers are a dime a dozen on the OSS world, so I'm largely unimpressed. It *is* nice, but not exactly "existential".

Quote:

But there's a living example to look at: AROS. AROS runs on top of Linux, yet it is (or at least can be) totally compatible with AmigaOS at source level (and binary level on the right architectures). AROS could be improved in such a way as to make it transparent to the user that there's Linux underneath, and Linux programs could be integrated with AROS so that you could run them from the AROS CLI and their graphical output would be shown on the AROS screen(s).

Ah but you see, there's the thing: AROS on top of Linux is just a software layer. Linux is a fully-fledged OS, so AROS doesn't add anything other than a software compatibility layer, basically "syntactic sugar" (AROS on the bare metal, on the other hand, provides necessary OS services that wouldn't exist without it). That's basically my point: think back to Windows 95. Windows 95 was just a desktop environment and the Win32 API on top of MS-DOS. Windows 95 added a lot of "value" over MS-DOS because DOS is basically just a program loader and device driver layer. Linux, on the other hand, is a fully-fledged OS, and (once you've added X11) putting anything on top of it doesn't *really* gain you anything that isn't cosmetic. Same thing with MacOS X and BSD (now I've done it - I've compared OSX to Win95. Boy am I glad this isn't Digg )

@Carl-S

Oh my gosh, here I am conversing with one of my Amiga heroes. I really never
expected to see this day.

I voted "I'm not sure". I certainly agree with most of what you have written,
but with some reservations.

Complete computer system, definitely. I'm not a hardware person, so when I
bought my AmigaOne I had Forefront actually build the hardware. It has given
me no problems at all and that is the way I like it. If I'd have tried to do
it, there would have been a huge mess. AOS4 wasn't out yet, so I did have to
install it. Being an Amiga, the installation was simple. However, most people
are not going to be able to install themselves. Definitely complete system.

The part about enabling and creative people (or even people who aren't quite
so creative, like me) is also a no-brainer. As to the "millions",certainly I
am all for it, but how. Obviously, this is where low cost comes in.

My first thought was that I never had thought of Amiga as low cost. My A2000
with GVP accelerator/ram/SCSI/HD was certaibly not low cost. However, upon
reflection I realized that actually my first Amiga was the A500, and I guess
that being low cost really was a factor in my purchase of it.

The (low cost) devil is in the details, of course. I'm not exactly against
using the X86 CPU, it really is the operating system that defines the Amiga
for me. However, AOS4 is not written for X86. Do we want to see Hyperion and
the rest of the team go back and redo everything, with another long delay.
In addition, my understanding is that they don't have a license for X86, and
Amiga Inc. doesn't seem to have been very forthcoming about anything.

Another factor that enters into my thinking is that none of our applications
are written for the X86 CPU. Many of our applications are somewhat dated but
they are all that we have. Without Scion for my genealogy, I'd be lost. And
I couldn't operate without FinalWriter on my computer, as I despise MSWord
and OpenOffice. There would have to be some sort of emulation, I suppose,
but my experience with WinUAE has not been very satisfactory. It sure doesn't
feel like an Amiga to me. And if I have to spend half of my time running in
emulation, what is the point of even calling it an Amiga (much as I hardly
even think of my W2K machine running WinUAE as a Windows computer anymore).

I don't know if you have had the opportunity to use AOS4, Carl, but I think
you'd be pleased with it. Even at Update4, it is about as much Amiga as it is
possible to be. You'd know it was your "baby" right away. I don't want to see
the AOS4 path abandoned for any possible future. I won't believe in the AOS5
that Amiga Inc. has touted until I see it. I'd like to see AOS4 continued and
available on hardware for everyone. Then we can begin to think about future
plans for porting to other hardware -- or any other possible paths.

One parting thought is that my feeling is that at the moment an even larger
need is for applications. Amiga applications. Find some way to get printer
and scanner drivers, and get TurboPrint updated. Get AWeb & iBrowse updated.
These are areas where we are really lacking these days. For this we need
hardware for developers, and we need it soon.

In summary, I agree with your definition, but how do we accomplish it. The
fact is that it is pretty hard for any new computer to succeed against MS.
They have the market pretty much to themselves, I'm afraid. Looking at the
situation as it exists, I think the Amiga is probably going to remain a
hobbyist system. I hope I'm wrong. Maybe you have some answers -- some
thoughts about how this can be accomplished. I welcome them.

It was a delight to have you write to us, Carl. Please pay us a visit now
and again.

Regards,
Don Feldbruegge

Sorry, Double post.

Last edited by DonF on 25-Oct-2006 at 04:49 PM.

Triple post. Sorry again.

Last edited by DonF on 25-Oct-2006 at 04:54 PM.

Goldarn it, quadruple post. I really apologize.

Last edited by DonF on 25-Oct-2006 at 05:00 PM.

@filmamigo

Quote:

I disagree -- in Canada, at least.

Looks like Commodore marketed very differently in different countries then (heh, Commodore screwing up, whoda thunk?) In South Africa, when I bought my first Amiga, an A1200 cost about the same as a 40MHz 386DX with VGA, Sound Blaster, hard drive (I forget the size) and Windows 3.0. In terms of CPU, the 40MHz 386 owned the 14MHz '020 in my Amiga, but the custom chips in the amiga generally made up for it - until you decided to do something CPU intensive like raytrace. The A1200 also only had 2MB and the 386 came with 4MB, but that was OK because AmigaOS programs tended to be smaller than PC programs. Still, when you looked at the bare hardware, you got more for your money with the PC, but the Amiga won out because Windows 3.0 is a turd.

@CodeSmith
Quote:

Cool APIs to do just about anything and neat window managers are a dime a dozen on the OSS world, so I'm largely unimpressed. It *is* nice, but not exactly "existential".

You're absolutely right. OSX is lovely to use - Everything is simple, most things are very easy to find and everything just works. That's not the case with some of the OSS GUIs - though they are getting much, much better. They'll catch up, given time.

Last edited by gnarly on 25-Oct-2006 at 04:21 PM.

_________________

@CodeSmith

Quote:

Not really, no. Maybe it's because I'm not a member of the Mac community and see Macs as just computers, but as I said in another post, Macs are OK but nothing special. Maybe it's because I've used Linux for years, and so I can see through Apple's marketing to see what MacOS X really is: a set of libraries and a neat window manager on top of a Unix clone. Cool APIs to do just about anything and neat window managers are a dime a dozen on the OSS world, so I'm largely unimpressed. It *is* nice, but not exactly "existential".

We're down to personal tastes here, I'm afraid, and therefore the discussion is not on an objective ground anymore. If you prefer BSD over MacOSX, fine, but then I could say that a kernel is nothing more than a scheduler, a bunch of device drivers and some "cool" memory handling routines.

Gives a completely different view to the thing, doesn't it?

Quote:

Ah but you see, there's the thing: AROS on top of Linux is just a software layer. Linux is a fully-fledged OS, so AROS doesn't add anything other than a software compatibility layer, basically "syntactic sugar" (AROS on the bare metal, on the other hand, provides necessary OS services that wouldn't exist without it).

And that's where you're totally wrong. AROS hosted and AROS native are different only at the level of the device drivers, anything else is exactly the same. If you are going to tell me that what matters more in an OS are device drivers, I'm afraid I'll have to strongly disagree.

You're perhaps confused by the fact that Linux is more advanced than AROS, but the point of making AROS run on top of Linux is to merge both the functionalities of AROS and Linux, so to get a system that is better than both taken as single items.

And Linux is more advanced than AmigaOS4 as well, and will be so for the forseable future, that's why it would make sense to make AmigaOS4 run on top of it.

Quote:

That's basically my point: think back to Windows 95. Windows 95 was just a desktop environment and the Win32 API on top of MS-DOS. Windows 95 added a lot of "value" over MS-DOS because DOS is basically just a program loader and device driver layer.

I really don't get all those "just"'s of yours. "Just" and "Win32 API" really don't fit well together. As you say it, the DOS was just a loader and a device driver layer (but device drivers could be written inside Windows itself as well), is that really the most important part of an operating system? If so, why isn't everyone stuck with DOS still? Hope you're not going to tell me that the reason for that is that people are so much after "fancy" stuff.

Quote:

Linux, on the other hand, is a fully-fledged OS, and (once you've added X11) putting anything on top of it doesn't *really* gain you anything that isn't cosmetic. Same thing with MacOS X and BSD (now I've done it - I've compared OSX to Win95. Boy am I glad this isn't Digg )

Oh well... if you think OSX adds just "cosmetics" to BSD, who am I to disagree?

_________________
“It is well enough that people of the nation do not understand our banking and monetary
system, for if they did, I believe there would be a revolution before tomorrow morning.”
~~ Henry Ford

@abalaban

Quote:

abalaban wrote: @falemagn so what the advantage of being hosted for AROS ? Except the fact that it can't crash the whole machine (beeing launched as a task of the underlying OS) and that it does not have to boot the machine.

The main advantage is that it has access to all device drivers of the underlying host operating system, benefiting from improvements made to it by third parties. You also get virtual memory (in the sense of memory being swapped to disk) for free, without AROS knowing anything about virtual memory per se. And, from the developer's point of view, you get the ability to do post-mortem analysys which give you instantly the place and reason of the crash.

If AROS hosted had to be improved, you'd also get the ability to run applications available to the hosted environment as if they were native to AROS, retaining their most wanted properties like the ability to make use of virtual memory and memory protection.

_________________
“It is well enough that people of the nation do not understand our banking and monetary
system, for if they did, I believe there would be a revolution before tomorrow morning.”
~~ Henry Ford

@falemagn

Quote:

And that's where you're totally wrong. AROS hosted and AROS native are different only at the level of the device drivers, anything else is exactly the same. If you are going to tell me that what matters more in an OS are device drivers, I'm afraid I'll have to strongly disagree. You're perhaps confused by the fact that Linux is more advanced than AROS, but the point of making AROS run on top of Linux is to merge both the functionalities of AROS and Linux, so to get a system that is better than both taken as single items.

I'm not confused by the fact that Linux is more advanced than AROS. AROS on Linux seems to be serving the exact same purpose as WINE the Windows API layer. People use WINE to run those Windows programs they want/need to, while still running Linux. If I want to run AmigaOS applications on Linux, I use UAE, which comes with chipset and 68K CPU emulation built in. There are no Amiga x86 "killer apps", so again, as an end user I see no benefit to "AROS on Linux". As a developer, sure, I can understand someone preferring the Amiga APIs over the Linux ones, but a healthy platform has a lot more users than developers. All this would apply to a version of OS4 running on Linux instead of Exec too.

Last edited by CodeSmith on 25-Oct-2006 at 05:49 PM.

@Carl-S

I see a lot of good discussion here...

I have scanned the responses. I do not have time to read every detail. A wide range of opinions have been expressed.

Over 25 years of creating systems, software, and products I've found two important patterns. You can either converge, or you can diverge. With the former, after a period of time, you end up with something concrete and useful. With the latter, after a period of time, you have nothing useful. Yes, that pattern is a broad generality, but I've seen it proven time and again.

When I evaluate working on projects with a company, org, or even other people/teams, I try to determine what pattern they follow. I want the effort to yield something concrete. I don't want to waste my time. (Note, that I do not apply that to forums... This is an open community for free expression. You are free to diverge all you want.)

The main point of my message boils down to this: find a unity of purpose.

I would never claim that I have a perfect solution for Amiga. Perfection is in the eyes of the beholder. But, I do know what I want from computers (and computing).

-Carl Sassenrath
OS Manager, Commodore Amiga
Advanced Technology, Apple Computer
Founder, REBOL Technologies

PS: I'll come back every few days and scan the messages.

Last edited by Carl-S on 25-Oct-2006 at 06:03 PM.

_________________
When replying, please keep it short and sweet, and I will read it. If it is long, I read only the first few lines to decide if I want to read the rest.

@falemagn

Quote:

Probably "random" wasn't the right term to use, but I used it to give the sense of those quotes being kind of out of the proper context. Carl has given a definition. A definition, by - err- definition, doesn't require more words in order to be understood and convey its message. If it does, it's not a definition. Carl has willingly separated the definition of "Amiga" from his own personal preference about how he sees that definition could be applied to the current situation. He clearly prefers AmigaOS4 over MOS and AROS, but then he surely prefers RebolOS over AmigaOS4. Taken in this context, AmigaOS4 becomes kind of irrelevant because it's just not required in order to understand the definition of "Amiga" that Carl has given. If Carl thought that AmigaOS4 were relevant to the "Amiga" definition, he'd have put it in the definition.

1. falemagn in post #110 says that AOS4 does not enter the picture of Carl's Amiga definition/future vision (heh... we have a "definvision" );
2. saimo, in post #112 just wants to say that AOS4 does enter the picture, although as a not strictly required component, because Carl-S himself brings it in in its definvision - BTW, in case you missed it, the *whole* post constitutes the definvision;
3. falemagn, in post #115 insists that AOS4 is not a prerequisite, making it look like saimo says that AOS4 is strictly necessary, whereas he had already clearly stated the opposite;
4. saimo, in post #119 is forced yet again to defend himself and make it very clear that he never said that AOS4 is a prerequisite;
5. falemagn, in the post I'm aswering here, starts another run of the circle by saying that AOS4 is irrelevant as it does not enter the definition...

... do you want to go on and on and on forever? In my first reply (with not many more words than these), I basically reported Carl's words by saying: AOS4 does not define "Amiga", but does enter the picture when talking about making that "Amiga" return (which is what was being discussed). That's so simple and clear I'm surprised it originated a discussion.

Quote:

I, again, think you're taking Carl's words out of their proper context. When Carl talked about Linux, he surely had in mind not just the kernel, but the whole operating system, which sees itself deployed in the multitude of distributions available today. This is the same thing I myself I talked about in a earlier post in this thread. Linux itself, the kernel, doesn't suffer from any particular problems, the problem is what surrounds the kernel. This is something that should be taken in high consideration when talking about these issues. Whether or not a future AmigaOS would use Linux as a kernel has certainly no bear on the usability, agility, flexibility and lightweightness of the system as a whole.

Since what we are discussing about is AROS on top of Linux with the capability of executing Linux applications, the kernel you are talking about is not what just executes task switching and provides HW drivers: it's a full-fledged Linux minus the desktop environment (Gnome, KDE, whatever). That *has* a bear on agility/flexibility/simplicity and *has* a different nature from "Amiga".

Quote:

To make a very simple example, consider Amithlon: Amithlon is just that, AmigaOS on top of Linux. You'd be hard pressed to find anyone claiming that the Amithlon experience is neither agile or simple.

Wrong example: does Amithlon allow to run Linux applications like you were proposing (which is what I replied to)? AFAIK, it does not.
Moreover: does taking care of Amithlon mean taking care of both the Linux and the AOS side? I guess it does, and this alone constitutes a more complicated system that a plain Amiga-ish system.

The AROS+Linux system you are talking about has added complexity that makes it more complicated, heavier and different in nature than an Amiga-only system.

Quote:

Quote: Let's not forget that this quote was intended to be as a reply to just the AROS+Linux and simplicity/agility discussion, so I'll stick to just that even now. That's basically another way of saying that you don't want to discuss things in the context they belong to.

No. That's basically a way of saying that I was discussing that particular subject (AROS+Linux, as defined in your post #110) and not the rest (because I could not follow the whole thread).

Quote:

What's the use of discussing just one particular subset of the argumentations I put forward, considering that they heavily rely on the rest of the argumentions which you willingly avoid to consider?

First off, who told you that I "willingly avoid to consider"? What do you know about my spare time, my health conditions, etc.?
Then, in the first post I replied (post #110 again), you introduced a precise system and I expressed my opinion about it... my reasoning was intended about just that precise system, and not meant to counter any other cosideration of yours (which I did/do not know about), so it was perfectly legitimate and, as forum discussions can be, useful.

Quote:

But you still haven't shown in which way the AROS+Linux architecture is not agile nor simple.

In my very first reply I did not just say "AROS+Linux is not agile nor simple because I say so". I said (as reported also above with different words):
Quote:

The AROS layer (with its own issues) on top of Linux (with its own issues) constitutes a complicated system with lots of issues... so, where's Amiga's simplicity and agility?

I think that the assumption that glueing together AROS and Linux is likely to yield a more more complicated system than a plain AOS (especially considering that the Linux part alone is already rather complicated in itself) is quite fair.

saimo

_________________
RETREAM - retro dreams for Amiga, Commodore 64 and PC

@thread

On the notion of a prospective customer standing in WalMart, looking at a boxed Amiga and wondering if it is the right machine for him to use straight out of the box... and speaking of the prospect of not having Windoze apps on the new Amiga, that might possibly be solved by a package called ReactOS that runs Windoze apps without having to install Windoze on a machine, and it's free, but it's also pretty much in alpha. On the other hand, there is a PowerPC version in the works.

Maybe we can add something like this, when the time comes, to an Extras disc, or to a pack of freeby discs, included in the box with the machine.

_________________
"I love peacenicks, they're so easy to conquer." --Ivan J Ironfist, the Dictator

@ regarding the cost of the Amiga

In the UK:

A500 was £399.99 when it sold most (end of 1989-end of 2000) - compared with £1500-2000 for a monochrome soundless PC.

A1200 was £299.99-399.99 in its base form - compared with £1500-2000 for a VGA PC with a basic soundblaster.

Dont forget that graphics card and souncard development didn't take off until Doom made its appearance...

_________________
Sven Harvey
Amiga Mart in Micro Mart, Geekology 4M@, and other places
A1000, A2000, A1500 A500, CDTV, A500+, A600, A4000, A1200, CD32, AT A1200HD, A1-XE

@saimo

Saimo,

no point in me answering any further to your posts, as long as you keep that kind of arrogant attitude. It's clear that none of us is going to change his mind anyway, so let's stop it here.

_________________
“It is well enough that people of the nation do not understand our banking and monetary
system, for if they did, I believe there would be a revolution before tomorrow morning.”
~~ Henry Ford