@Lou

Quote:

Dongle? That reminds me of a joke. What's that useless piece of skin attached to the [insert term for your choice of male or female reproductive anatimy here]?

Your manhood!

Quote:

I don't see any reason why the OS4 task based approach couldn't go on. Tasks could run as they do now as tasks and a separate stack of 68k libraries could be used to protect the newer OS4 system.

Look: It's been 16 (!) years since the last release from Commodore. Since then, numerous developers has been hacking around with the OS, trying to patch it, improve it, etc... I think, but I may be wrong, that if adding memory protection was as easy as you say, it would have been done. And all three (AROS, MorphOS, OS4) would already have it since some time... don't you think ?

_________________
http://www.warpdesign.fr/

@Hypex

it was a virtual machine running in rootless mode.
I don't see how that's silly. Compatibility was amazing.
---

Also I don't know where you got that bogus info about macos using 8 bits of a pointer to store info.

Last edited by Frek on 25-Jul-2009 at 02:35 PM.

@Leo

Quote:

Amiga still has it. Because any program requires it... I'm sorry, but that's far more of a problem than keeping "--" to specify command line arguments ;)

I don't know why they didn't better protect it and put in a more safer API. I thought OS4 was the point of that, to update the API with a new one so these things didn't happen.

I mean, AFAIK there is already a limit to how much memory 68k apps can use. I agree it is annoying how easy it is to bring down the OS, or have the mouse crash, leaving the only option to reset. But I thought the newer OS4 programs would be safe in this regard or even be running in a safe enviroment where this sort of things weren't allowed to happen.

As an example, one of my programs crashed with a NULL pointer access in the OS. The OS was updated to strictly adhere to the autodocs. Now I was unaware I passed the OS a NULL pointer. So it was good I found out. But it is bad if the "new" OS didn't check the pointer because the docs say it mustn't be zero! No wonder it is unstable, it trusts the programmer too much!

Now we have talk of another more robust OS4 in the works. And see things are working towards this but why wasn't this clean breaked from the start? And the older API just kept for 68k and managed by the OS?

@Leo

Quote:

I think, but I may be wrong, that if adding memory protection was as easy as you say, it would have been done.

This isn't specifically about memory protection, I was talking about running 68k apps on the system. I don't see why each task would need it's own AOS sandbox or UAE like environment when all the 68k apps could run together and an OS3 68k API could be sand boxed.

Quote:

This isn't specifically about memory protection, I was talking about running 68k apps on the system. I don't see why each task would need it's own AOS sandbox or UAE like environment when all the 68k apps could run together and an OS3 68k API could be sand boxed.

It could have been if it had been thought this way from the start (like MorphOS for example). The thing is that OS4 wasn't supposed to last 5 years (or so). It was supposed to last maybe one or two years. And later, a sandbox could have been created. Now it's possible, the thing is we are 7 years later...

Last edited by Leo on 25-Jul-2009 at 03:17 PM.

_________________
http://www.warpdesign.fr/

Quote:

Now we have talk of another more robust OS4 in the works. And see things are working towards this but why wasn't this clean breaked from the start? And the older API just kept for 68k and managed by the OS?

Ask them... ;)

Last edited by Leo on 25-Jul-2009 at 03:19 PM.

_________________
http://www.warpdesign.fr/

@Frek

Quote:

I don't see how that's silly. Compatibility was amazing.

It wasn't the compatibility. It was that it had to boot an OS inside the OS. Okay when it's done but it slowed the process down and just seemed silly the need to boot an OS to run a program, especially in light of the fact that Classic Carbon apps ran straight out of the box. And they were in the old Mac format and using an old method to access the API.

Quote:

Also I don't know where you got that bogus info about macos using 8 bits of a pointer to store info.

It was from a well respected AmigaOS developer. With 24-bits IIRC you have a 16MB address space. Beyond and things break. This could also be the reason that A500 programs broke on OS2. It's also why Microsoft AmigaBASIC broke on a 68020. That used to mask every address it looked at to 24-bits! What the?! That explained why trying to load a program that was in the current directory and being told the file wasn't found because it didn't exist, didn't work!

@Hypex

Well this "respected AmigaOS developer" is a dumbass if he claims that; It's not true.

Regarding Classic, as I said it's a virtual machine, it's not a API reimplementation, you think it's silly to see AmigaOS boot in UAE too?

I hope it's ARM. I really see it as the way to get a user base of > 2000 people in the world. Looking at the pics on this site, Amiga does have a good look. I could see it being used a lot.

@Hypex

Quote:

AmigaOS also has some pointer-typing. The highest bit is used as a flag in relation to memory handling, ergo, it has a maximum range of 2GB.

Sorry, Damien, but that is utter rubbish.

The only limitation is that there are still some OS components (and third-party apps) that have not been cleaned of "int32 addrPtr" type declarations. Since some of them are actually in the API, we have had to leave them in place, rather than change the API and break things.

_________________
cheers
tony

Hyperion Support Forum: http://forum.hyperion-entertainment.biz/index.php

@EVERYONE

Why can't people just shut up, sit back and wait to see and hear.

The problem with everyone here is that everyone is building up expectations that are guaranteed to disappoint everyone except the few that gets it right. This is because you are putting too much hope and energy in YOUR vision and ideas then what may actually occur because we have A) Way too many people with ideas and Hyperion - hell not even Microsoft can possibly fulfill all your ideas.

A few or none of you may be right. The rest of you will not be right. The point is, quit hypothesizing what it might be. It is like a child hypothesizing about what is in the Christmas present. The problem with hypothesizing is by human nature - you begin to want and expect it and when you find out it is different then what you worked yourself up for, you end up being disappoint because you easily aim too high for a company of so small.

I don't hope or put any level of expectation or desire, That way, whatever it is, it is better then what I hoped for. Then it makes me feel happier then feeling disappointed.

Have a good day.

Quote:

It is like a child hypothesizing about what is in the Christmas present.

It's also childish giving some hint about your bigger project and not letting the community know months after... Not saying it's not professional either.

Not letting your community know what happens is the worst thing you can do as a company.

- They make the products, you are right. And neither them nor Microsoft could fulfill all ideas (not to mention a lot of ideas are simply not realistic in a technical point of view, or in a business point of view).

- But *we* make the community by buying their products...

EOF.

Last edited by Leo on 26-Jul-2009 at 09:34 AM.

_________________
http://www.warpdesign.fr/

@Wildstar128

Quote:

Wildstar128 wrote: @EVERYONE Why can't people just shut up, sit back and wait to see and hear.

because this wild speculations are fun.

_________________
AmigaONE. Haha. Just because you can put label on it does not make it Amiga.

I borrowed this comments from here (#27 & #28):
http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=38873&forum=2&start=20&order=0

@tonyw

Quote:

The only limitation is that there are still some OS components (and third-party apps) that have not been cleaned of "int32 addrPtr" type declarations. Since some of them are actually in the API, we have had to leave them in place, rather than change the API and break things.

Its impossible to change that unless you switched to 64bit CPU architecture (or unless you are using unions of 32bit).
Anyway the device and library bases addresses has to be whit in 32bit address space to be seen by other 32bit applications,

If applications used privet / virtual memory, then that will free up shared address space, whit the 32bit bit (4Gb) range.

I find it bit strange that we are talking about this, memory usage is usually between 64mb to 128mb in usage.

Last edited by NutsAboutAmiga on 26-Jul-2009 at 03:41 PM.

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

@Leo

Quote:

Leo wrote: Quote: It is like a child hypothesizing about what is in the Christmas present. It's also childish giving some hint about your bigger project and not letting the community know months after... Not saying it's not professional either. Not letting your community know what happens is the worst thing you can do as a company. - They make the products, you are right. And neither them nor Microsoft could fulfill all ideas (not to mention a lot of ideas are simply not realistic in a technical point of view, or in a business point of view). - But *we* make the community by buying their products... EOF.

I can understand that point too but I hate to see the community build up too much hopes for something and when it does come - it probably isn't what most if not all of what everyone is hoping for and is disappointed. Don't ruin the surprise for yourself and the joy of a surprise.

That is why I am trying to pop the dream bubbles and bring everyone back down to earth and not expect SO much. It can be fun, like Arko said, but it is dangerous and you can hurt your joy about something.

Some ideas are unrealistic but people build their hopes for it, though. When you build your hopes for something unrealistic then you be disappointed if it falls short of unrealistic even tough it is as far advance as realistically can be achieved.

Hopes & Expectations doesn't always follows logic and sensible.

@Wildstar128

They have a good product. I hope for their sake that they can introduce it to more people by putting it onto ARM.

@Al4

Quote:

Al4 wrote: @Wildstar128 They have a good product.

don't be kidding, Hyperion is fears an open market.

Quote:

I hope for their sake that they can introduce it to more people by putting it onto ARM.

Or better to some available used Macs that are faster, better available and cheaper.

_________________
AmigaONE. Haha. Just because you can put label on it does not make it Amiga.

I borrowed this comments from here (#27 & #28):
http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=38873&forum=2&start=20&order=0

@Leo

Quote:

It could have been if it had been thought this way from the start (like MorphOS for example).

As I remember they chose it this way on purpose so there was no sandbox.

Quote:

The thing is that OS4 wasn't supposed to last 5 years (or so). It was supposed to last maybe one or two years.

It wasn't! I thought OS4 was IT!

Quote:

And later, a sandbox could have been created. Now it's possible, the thing is we are 7 years later...

The task based emulation was a feature of OS4. And personally I liked it. But the silly thing is there has been talk of a sandbox for OS4 native apps as they are now, what the?!

@Hypex

There are no current developments moving to a sandbox.

With respect to the project (which is proceeding nicely), few real Amiga fans will be disappointed, quite the contrary.

But as always, the proof of the pudding is in the eating.

_________________