Wither Natami?

[HenryCase]

the_leander wrote:
Haiku, like BeOS took some of the best ideas from AmigaOS, things like the datatypes system and brought it bang up to date with things like video codecs etc.

Apart from datatypes, in what ways is BeOS/Haiku similar to AmigaOS?

bloodline wrote:
Hans... no offence meant here, but have you ever done any Amiga development at all?

No offence, but... do you intend to stir trouble or are you just naive? In any case, just so Hans_ doesn't have to rise to your bait himself, he is working on OpenGL stuff for AmigaOS4 IIRC.

@all
I'm sure you're all aware of CAOS, the operating system the Amiga was destined to have from the start. Here's some information about it if you need a refresher:
http://www.thule.no/haynie/caos.html
Now from what I can tell from that synopsis CAOS was very similar in design to what we now know as AmigaOS, but with a few key differences that allowed for better MP, better file system, etc...

Two questions:
1. Will copying or adapting some of the design features from CAOS allow us to move into a more stable AmigaOS without having to rewrite all the subsystems from scratch?
2. Will these CAOS features restrict us in building the application sandbox Hans_ is proposing?

[bloodline]

HenryCase wrote:

bloodline wrote:
Hans... no offence meant here, but have you ever done any Amiga development at all?

No offence, but... do you intend to stir trouble or are you just naive?

Don't be an idiot! That was a legitimate question, Hans is asking questions that honestly give the impression that he has had very little experience messing around with AmigaOS.

In any case, just so Hans_ doesn't have to rise to your bait himself, he is working on OpenGL stuff for AmigaOS4 IIRC.

:-? And? What bait? What "OpenGL stuff"?

@all
I'm sure you're all aware of CAOS, the operating system the Amiga was destined to have from the start.

I'm very aware of CAOS... and there is a big reason why we don't have it now.

Here's some information about it if you need a refresher:
http://www.thule.no/haynie/caos.html
Now from what I can tell from that synopsis CAOS was very similar in design to what we now know as AmigaOS, but with a few key differences that allowed for better MP, better file system, etc...

CAOS was totally different from AmigaOS. And guess what it didn't work and it couldn't be made to work in budget or in time.

Two questions:
1. Will copying or adapting some of the design features from CAOS allow us to move into a more stable AmigaOS without having to rewrite all the subsystems from scratch?

CAOS was a fundamentally different system built around our beloved exec. Oh and it didn't work... if you read what is left of the specs, you can see it was an odd system... probably even less compatible with the all pervasive POSIX than what we have now...

2. Will these CAOS features restrict us in building the application sandbox Hans_ is proposing?

Forget about the sandbox... ok?

There is no future with AmigaOS anymore... we need to make the best of what we have, and develop that to the best of the concepts strengths... rather than these silly pipe dreams, which serve to do little more than make our favourite OS look really stupid.

[HenryCase]

bloodline wrote:
That was a legitimate question

Maybe, but what did you intend to achieve by asking it?

bloodline wrote:
What "OpenGL stuff"?

Adding functions/features to TinyGL.

bloodline wrote:
CAOS was totally different from AmigaOS. And guess what it didn't work and it couldn't be made to work in budget or in time.

Those reasons were financial rather than technical, as far as I can tell. See this quote from Carl Sassenrath: "CAOS was contracted out, for the most part, to a company that felt Unix was a better choice and didn't buy into my design. They became history when they started using their Sun development systems for other projects, not the Amiga higher level OS functions."

bloodline wrote:
CAOS was a fundamentally different system built around our beloved exec. Oh and it didn't work...

...because the coding wasn't finished. That doesn't mean we can't apply some of the ideas from CAOS to our newest AmigaOSs.

bloodline wrote:
if you read what is left of the specs, you can see it was an odd system... probably even less compatible with the all pervasive POSIX than what we have now...

Not everyone sees POSIX as a good thing. Seems to me the more you strive for total POSIX compatibility the more you turn your OS into a UNIX variant. As long as apps aren't super difficult to port why care about POSIX?

bloodline wrote:
rather than these silly pipe dreams, which serve to do little more than make our favourite OS look really stupid.

This goes back to the can it be done/should it be done split I outlined in our last MP discussion. Should it be done? Probably not. Can it be done? That's the interesting part. Forget about logistics, discuss the technical details.

[SamuraiCrow]

Hosted AROS runs AROS software in a sandbox on top of Linux.  If you don't want the sandbox and want to run it memory-protected, use Linux itself.

[the_leander]

HenryCase wrote:

the_leander wrote:
Haiku, like BeOS took some of the best ideas from AmigaOS, things like the datatypes system and brought it bang up to date with things like video codecs etc.

Apart from datatypes, in what ways is BeOS/Haiku similar to AmigaOS?

*dusts off his cope of CU Amiga, dated July 1998...

Client-server internal architecture, multithreaded throughout, Pervasive multithreading with realtime multitasking with an object oriented API that is clear and simple.

The IO system is modular, dynamically loaded (just as with .library and .device)

That said, there are some major differences, it has memory protection, virtual memory, the graphics system is modular to the point where a driver for a graphics card can be installed and used without rebooting.

The way it handles filetypes is different, although newicons and deficons aparently have similar capabilities.

The filesystem is journaled and offers realtime searches that to this day is very powerful in comparason to OSX's and Vistas.

It also has posix compliance and a built in TCP/IP stack.

The big difference though was the fact that API forced you to write your application in a multithreaded fashion, which for a complex application wasn't such a bad thing, for a simple application (such as unzip, for instance) it was less useful, indeed it could be a disadvantage.

Haiku (the last time I checked) was implimenting full Multiuser support (the option was always there but never implimented in the original BeOS). Zeta 1.5b had fully functional Multiuser support however it's development was cut short when the guy who started Zeta was found to have lied about his having a licence with Palm for the original BeOS code.

[Piru]

@Hans_

Could you please provide a solid example so that I understand what the problem is?

selib.c

Constructs such as this are extremely common, and WILL die horribly with MP.

I don't think that a manual check would be needed for every occurrence.

Why not?

[Piru]

@HenryCase

1. Will copying or adapting some of the design features from CAOS allow us to move into a more stable AmigaOS without having to rewrite all the subsystems from scratch?

No.

[Piru]

@SamuraiCrow

Hosted AROS runs AROS software in a sandbox on top of Linux. If you don't want the sandbox and want to run it memory-protected, use Linux itself.

Examundo. You hit the proverbial nail in the proverbial head.

This is the exact reason why such NG-AmigaOS-project makes absolutely no sense at all. You'd only end up with 0 applications, especially if you wouldn't be POSIXy (apparently some see it as a bad thing...). Also, lets not forget the horrible pain that'd ensue from having to write drivers for the tens of thousands of x86 devices. Either do that, or severely limit the possible setups for the OS.

I'd take the negligible overhead of running hosted on top of the Linux kernel any day.

[amigaksi]

>This week a cascading failure beginning in the A2091 lead to the death of the card, a CD burner and hard drive.
...
>I'll buy the first clone that can reliable meet or exceed the capabilities of my old A2000/040/32MB RAM.

How does the hard drive controller card failing relate to the A2000 motherboard?  Won't just replacing the A2091 card with GVP SCSI card help?  I have had a GVP SCSI for many years and has not yet failed although the hard drive attached has failed since it's well past its MTBF; however, it still works if I press On/Off a few times and get it to spin up.  The only motherboard that ever failed me was the one that had the battery leak.

[amigaksi]

by asymetrix on 2008/8/3 22:45:02

>This thread is so silly.

Well, your comments don't help to improve it.

>1) This shrinkage changed the speed of the original chips by laws of physics.
>This equals not cycle exact.

First you have to have a frequency that you are talking about before you claim it's not cycle exact.  You can have something execute by the time the cycle time you want is expired, and the shrinkage won't affect the cycle exact nature as everything will be ready to be clocked in at the next cycle.

>2) from OCS to AGA the design changed again, speeds changed again, design changed for 32 bit support.

>This equals cycle change again.

No, the OCS Copper still run at 3.57954525Mhz on AGA.  The OCS Paula still has a sampling rate divisor of 3.57954525Mhz.  The CIA timer still runs at 715909Hertz.  CPUs were purposely put at higher clock rates with faster memory; they weren't trying to keep them running at the same clock frequency since the timing could still be done via other methods (CIA, Copper, using a Paula sound channel as the original poster's sound software does).

>There is no such thing as cycle exact, its just marketing hype.

Yes there is and on the Atari 800, even the CPU at 1.7897Mhz is cycle exact as well as it's timers, DLI, VBI, etc. since they never bothered to upgrade CPU speeds.

>IF hardware can run Amiga OS and Amiga programs the same way as Commodore Amiga then its Amiga. Simple as that.

If the programs use the cycle exact nature of the Amiga which was maintained in all of its models and it does not run the same on a "new" software or hardware amiga, then that's not a real Amiga.  Why would Commodore bother keeping all those timings the same in future models if cycle exact was just hype.

>Software, hardware, simulation, emulation its all Amiga.

I believe many people familiar with hardware will disagree with you.  

>Simply put, do you actually think a normal user cares what hardware it is ?

You can also say fake diamond is the same as a real diamond if it looks the same.  But some people have use for the real one and to claim it's all the same is being vague and misleading.

[jj]

I mentioned this early.  Really do no see the point in running old OS stuff in a sandbox and the code a ccompletley new OS to run the sandbox in.  

We have already got a brillaint sandbox its called UAE, run it on any system you want (more or less). There are even linux installs around that are so small and with running UAE in mind, so you would not even not notice xit wasnt native OS3.9.

We already have MorphOS and OS4.  They are the most modern an amiga like OS is going to get.  

[bloodline]

HenryCase wrote:

bloodline wrote:
That was a legitimate question

Maybe, but what did you intend to achieve by asking it?

Your reaction to the question is far more interesting than either my question or Hans's answer could ever be!

I was interested in what experience Hans had, so that I could answer his questions better, perhaps he has experince in VxWorks... I don't know, but his questions were not those of someone who has spent many a bored evening messing around with an Assembler trying to get stuff to work...

You rection though, immediately makes me think there is far more to this than you are letting on, and not a good way. Someone is bullsh1ting us here!

bloodline wrote:
What "OpenGL stuff"?

Adding functions/features to TinyGL.

Bug fixing a library that is not native to the Amiga is hardly a way to learn about the internals of AmigaOS.

bloodline wrote:
CAOS was totally different from AmigaOS. And guess what it didn't work and it couldn't be made to work in budget or in time.

Those reasons were financial rather than technical, as far as I can tell.

During the development of a new Operating system, is there any difference... Any problem can be solved eventually if you throw enough money at it.

Technology is all about bringing in the features you want, on a budget and in timescale.

See this quote from Carl Sassenrath: "CAOS was contracted out, for the most part, to a company that felt Unix was a better choice and didn't buy into my design. They became history when they started using their Sun development systems for other projects, not the Amiga higher level OS functions."

See above.

bloodline wrote:
CAOS was a fundamentally different system built around our beloved exec. Oh and it didn't work...

...because the coding wasn't finished. That doesn't mean we can't apply some of the ideas from CAOS to our newest AmigaOSs.

It couldn't be finished. It wasn't possible to get it done in the time frame/budget that Commodore had, to get the Amiga out in time. If they had stuck with it and the project had rolled on for 2 more years... too much money would have been spent, and every other computer company would have had a chance to make something better. The Amiga would have failed before it even left the stable.

bloodline wrote:
if you read what is left of the specs, you can see it was an odd system... probably even less compatible with the all pervasive POSIX than what we have now...

Not everyone sees POSIX as a good thing. Seems to me the more you strive for total POSIX compatibility the more you turn your OS into a UNIX variant. As long as apps aren't super difficult to port why care about POSIX?

You have a choice... be POSIX compatible and have software or be some strange little curiosity OS... with no software.

All you really have now is Win32 or POSIX (ok, the MacOS X NeXTStep frameworks are gain ground, but Mac software really stays on the Mac). 99% of all free software is POSIX, and a new operating system needs free software to be in any way useful.

If you don't like POSIX, really what else is there?

bloodline wrote:
rather than these silly pipe dreams, which serve to do little more than make our favourite OS look really stupid.

This goes back to the can it be done/should it be done split I outlined in our last MP discussion. Should it be done? Probably not. Can it be done? That's the interesting part. Forget about logistics, discuss the technical details.

I'm happy to disscuss the technical issues, but I would prefer it if we were not just going around in circles... and as soon as an issue is almost put to bed as unsolveable... someone pipes up that I must be an Amiga hater, or that there must be a way for the amiga to make a massive come back, we just need tothink about it properly... both of which are untrue.

We are at the end of the developement, cycle here. The Amiga scene had some of the greatest minds available... Any problem that can be solved has been solved. AmigaOS is what it is. If you want to make a new OS go right ahead. But I like our little dinosaur.

[cicero790]

Hi all
I followed this and I was thinking. That if one changed the definition of where it must go to be modern? The string theory was canned and then rediscovered and is now hotter than ever. Bear with me for a second despite my low knowledge in this field. You can say that AROS is not VISTA, true.
AROS is not a Ferrari it is a salt lake sound barrier braking OS then run on a fast PC. It’s so fast it’s ludicrous. There was a guy yesterday that wanted to use AROS for some car application that will put  peoples life in the hands of an AMIGA OS. Samurai crow wrote over at the Amiga world, then talking about parallel hardware that MP was not necessary.  Is there another way to do it? Could it be a strength to not have a similar MP like all the dragons? Just a little comment, I may be completely wrong.

[bloodline]

cicero790 wrote:
Hi all
I followed this and I was thinking. That if one changed the definition of where it must go to be modern? The string theory was canned and then rediscovered and is now hotter than ever.

I don't think you know your Modern Physics very well... :-(

Bear with me for a second despite my low knowledge in this field. You can say that AROS is not VISTA, true.
AROS is not a Ferrari it is a salt lake sound barrier braking OS then run on a fast PC. It’s so fast it’s ludicrous.

But the AmigaOS design can't take advantage of modern hardware features... An OS that can use an 8core CPU... is going to outperform one that can only use one.

An OS that doesn't need to reboot every time there is a problem, is outperforming already one that does!

There was a guy yesterday that wanted to use AROS for some car application that will put  peoples life in the hands of an AMIGA OS.

As I understand it, his application is not mission critical.

Believe me, system that are mission critical... say used on aircraft, for example, put stablity and fault tolerence above all else.

You would be an idiot if you used AmigaOS in such a situation.

Samurai crow wrote over at the Amiga world, then talking about parallel hardware that MP was not necessary.  Is there another way to do it?

MP isn't necessary for SMP, but it's not very efficient if don't have VERY controlled memory management. Imagine if you have a CPU core trying to access memory that is on a separate bus, that is in use by another CPU... the speed would be terrible...

MP didn't wasn't invented to make AmigaOS redundant... it is a very useful feature.

Could it be a strength to not have a similar MP like all the dragons?

Singularity tried a different approach... M$ have dropped it now, so you can play with the source code yourself. It is very nice, and I look forward to seeing what the OpenSource community do with it.

Just a little comment, I may be completely wrong.

Indeed.

[cicero790]

You have a sharp sword Bloodline. It looks grim the way you present the situation.  But, can you see a way to adapt AROS to take advantage of multi core processors??

[bloodline]

cicero790 wrote:
You have a sharp sword Bloodline. It looks grim the way you present the situation.  But, can you see a way to adapt AROS to take advantage of multi core processors??

Yes, actually... This is an Area that Kal and Michal S on AROS-Exec.org can disscuss better :-)

I find concurrency issues very complex and it takes me a while to get my head around. In short it is possible, but the design of AmigaOS (and thus AROS) is not really very suited to SMP...