Amiga Inc have posted the list of changes for AmigaOS 4. The main point is that the Exec kernel, which is currently in 68k assembler, is being rewritten in C with major enhancements. Update: The Register is also covering the latest news regarding AmigaOS4 and other Amiga related products.
AmigaOS 4 Feature List Released
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Why do they say “It has been rewritten in C, as ExecSG, for AmigaOS 4. The following features are planned for ExecSG:…” Sounds to me like they’ve written it but they haven’t written it.
No wonder people are worried about vaporware – they make it sound like they haven’t really written anything. Just planned goals for something they haven’t really programmed yet. I hope that’s not the case but it sure sounds fishy…
Yes, Thomas Frieden of Hyperion (The AmigaOS4 project leader) has already completely re-implemented the new Exec SG kernel in the C language while adding some additional features to the original design.
I do agree though, that someone has mistakenly left out some important information as this message could now be interpreted wrongly, it should have said:
Exec is the kernel of the AmigaOS and is currently written in 68K assembly. Exec has been re-written in C and new functionality has been introduced to allow the deployment of OS 4.x on any suitable PPC hardware. The new Exec is called Exec SG (Second Generation).
The following features will be offered:
——–
From Ben Hermans on Amiga.org:
“What we are doing now is adding “virtual addressing”.
This was not planned originally but the benefits of this are so important that we decided to go ahead and do it.
Benefits include things like less memory fragmentation, faster and more efficient memory allocation and de-allocation, much more robust virtual memory implementation and finally automatic stack enlargement.
Between 30 to 50% of all crashes on the Amiga are due to programs running out of stack-space.
With OS 4, this won’t happen anymore as the stack can be automatically enlarged when the system detects a program is about to run out.”
I knew you’d have an answer for me.
They really should make the information on that page more clear, though.
My information is that the new Exec is written and running. Most of
the work has been done using the existing Amiga PPC accelerator cards.
There was extra work to do in writing a new BIOS for the AmigaOne
motherboard.
Most of the work has been done using the existing Amiga PPC accelerator cards.
Sorry but if the kernel will be/is being rewritten why not change the arch to x86?(64bit) ?
If Apple must leave the PPC why not Amiga too ?
Sorry but if the kernel will be/is being rewritten why not change the arch to x86?(64bit) ?
If Apple must leave the PPC why not Amiga too ?
First of all x86 64 is not even out yet (still Vapor ware) and Apple is not leaving PPC anytime soon
Why is being rewritten in C a good thing ?
Most of the AmigaOS was original in B or C but when
time got bye things were rewritten in assembler for speed reasons (it was the only OS which got faster with each
new version) But then intitial processor speeds were 7.16MHz later versions even 25MHZ with the 68030 and I even got an 50MHZ turbo board in my old A3000 🙂
Sounds kinda strange…what is an Amiga if the computer is running the OS with a C based kernel. The tweaked 68K assembly provided all of the speed and magic of the original machine…sounds like the Amiga OS is an OS without a platform…its coming to a PC near you…they need to do what Amiga and Commodore did in the past, buy a chip fab and start designing customized processors and chipsets…and optimize the OS to run on this new platform. This is the only way to bring the Amiga OS to a level that can surpass what is currently offered in PCs….PCs who’s specs are not to shabby at all…the only problem is that it takes real money to produce actual computer hardware…the Amiga platform is Dead.
Dano.
The Amiga was special because it was fast, simple and let you do amazing things.
People here have prejudged Amiga OS4 without ever seeing it. I have little to no doubt that if C compilers had been any good back then, that Amiga Inc would have used them. Writing in 68k ASM was the *only* way to get the speed that they needed. Writing in PPC ASM is a *world* of difference. Almost no one can hand code PPC asm as well as a good quality compiler can create it. There is just *way* too much to track. How many FP units are still running, how many integer units are still running, etc. You would almost have to write it on a spreadsheet or something. Ugh. And the quality of the compilers is *far* better today than 10 years ago.
As far as custom hardware, I would be very surprised if anyone out there could make custom chips that match the speed and quality of what goes into PCs today. Not that I am some PC shill, but be real – how many people does nVidia have doing video chips? You really think that Amiga could surpass that? At a similar price? How about sound? You can get sound chips that do what Paula (right? or was it denise?) did in 1985 for about $2, now. Or better for $10. Certainly the Amiga’s parallel port handling was very good. But parallel ports are *dead*. How about those quality 880k floppy disks? No one cares.
Be once said “It is dark inside the box”. If you are not writing ASM code yourself, or building a kernel, then it doesn’t matter in the least, really, what your processor is.
I want one. I HAVE TOO MANY COMPUTERS…
Well we have been waiting for 0S4 for two years now. To boot they are hitched to the powerpc, which is very slow, especially since all the fast PPC chips are unavailable to the Amigaone guys. They would have been better off hitching up to sparc.
>”Bill McEwen in summing up the events of the last year described as it
>hardest year he has ever had to face. However, he emphasized that they
would
>”deliver a brand new operating system for the desktop, new content for
>handheld devices, create a future and to keep growing”. Bill also stated
>that, “Here before Christmas I’m sure we’re going to see a lot of
AmigaOne’s
>under some Christmas trees”. ”
Note that last sentence. Now, if McEwen said it… let’s hope it happens,
eh? Any bets as to it’s accuracy?
Just so you know, I got this paragraph from Amiga Flame. It’s authenticity
(that Bill McEwen actually said it) has been confirmed by reply to me from
Gary Peake.
Yea!
Jared
Ty? Another Delay? The only date indication given by Hyperion, was that AmigaOS4 team is confident to deliver AmigaOS4 before the end of this year!
I believe AmigaOS4 should not be rushed. IMO the computing industry is still making its first baby steps. I am severely unimpressed by the inefficient, closed and bloated nature of both MacOS X and WindowsXP.
> Why is being rewritten in C a good thing?
AmigaOS will eventually become a multi-platform OS, by using only assembler they would have to start from scratch all over again for each processor platform. C is a portable high level language and thus takes far less effort to be ported across different processor based environments.
Although the OS will be more resource hungry as compared to 68k AmigaOS, it will still be very fast due to advances of modern computer technology. Sadly people have difficulty seeing that their hardware is extremely fast due to the inefficiency of modern operating systems.
Seeing Amithlon running on modern PC hardware could be a real eye-opener for most people. Although there is a big performance penalty while using 68k software emulated with Amithlon, the emulated 68k AmigaOS itself already runs circels around modern OSes which run natively on x86 hardware!
I tend to agree with Mr Bouma. Rushing things is good for no one. I still want an Amiga 4 os. I love the screenshots.
” I still want an Amiga 4 os. I love the screenshots.”
The screenshots are interesting so long as you remember that there is
still design work to be done and color schemes to be worked out. The
nasty color schemes on the current screen shots are just temporary –
and in any case, the users can set up their own.
One dyslexic user said he prefers red text on a cyan background, for
example – but it makes other people feel ill when they go near his
computer.
Dano posted
>Sounds kinda strange…what is an Amiga if the computer is running the OS with a C based kernel. The tweaked 68K assembly provided all of the speed and magic of the original machine…sounds like the Amiga OS is an OS without a platform…its coming to a PC near you…
The dependancy to old and obsolete HW was a fault in the long run. It was blocking the AmigaOS advancement.
According to the information that I read from some Hyperion guy’s interview, after they implemented the C -version of ExecSG, they have been doing a lot of optimizations for PPC in assembly. The C -language version is essential for future portability, but assembly is needed in many places to get maximum speed out of the HW.
Amiga is coming to a PC near you. You really should see how well Amithlon performs when it runs AmigaOS via emulation.
>they need to do what Amiga and Commodore did in the past, buy a chip fab and start designing customized processors and chipsets…
Totally impossible, there is not enough money (in the world) to do that in a decent way. Better use off the self parts.
>and optimize the OS to run on this new platform.
I think Amiga’s plan is to optimize the OS for a strightly limited number of HW platforms. On the rest of the platforms Amiga would exist on top of host OS (like Amithlon for AmigaOS on x86 and AmigaAnywhere for AmigaDE content on ARM/x86/etc.)
>This is the only way to bring the Amiga OS to a level that can surpass what is currently offered in PCs….PCs who’s specs are not to shabby at all…the only problem is that it takes real money to produce actual computer hardware…
It is not Amiga’s plan to surpass PC’s as a desktop computer.
If Amiga userbase starts to grow again someday, slowly, and developers start to come back to the Amiga desktop platform, then we might see also new custom Amiga HW for desktop, etc… But I would never bet my money on that, though.
>the Amiga platform is Dead.
You mean the old Classic Amiga HW platform, with custom chips … yes, pretty much so, there is no sense in developing the old HW any further.
(just note that OS4 will run also on those Classic Amigas that have PPC CPU (& radeons & soundblasters, etc…).)
AmigaOS is still a very useable OS for many computing tasks, with the OS4 update it only gets better.
> They really should make the information on that page
> more clear, though.
The feature list has now been updated at Amiga.com to make the page more clear with regard to Exec SG.
PPC native CD Filesystem
with limited Mount Rainier support (use your CDRW as a floppy)
IIRC, at first Mount Rainer wanted to use CD-RW, but the current specification is DVD+RW… I guess that’s what they meant by “limited”.
First of all x86 64 is not even out yet (still Vapor ware) and Apple is not leaving PPC anytime soon
Vapourware is defined by a “product” that doesn’t have any physical evidence. However, for a vapourware, there have been a awful amount of tradeshows showcasing Hammers, a awful amount of chipset makers and mobo makers announcing suport for Hammer, a very big company named Microsoft supporting it… Doesn’t sound like vapourware to me.
I tend to agree with Mr Bouma. Rushing things is good for no one. I still want an Amiga 4 os. I love the screenshots.
Me? I just hate it. The screenshots just look awful. I could build a Photoshop mockup of Amiga’s UI that look much more professional and less, well, ugly. Maybe it is appealing to old Amiga user, but…
The nasty color schemes on the current screen shots are just temporary
The colour scheme is fine. It is just the nasty background of the windows, as well as ugly looking icons.
“IIRC, at first Mount Rainer wanted to use CD-RW, but the current specification is DVD+RW… I guess that’s what they meant
by “limited”.”
It’s limited in that for the first version there will not be a UDF
file system, so basically the disks will only be readable on Amigas.
UDF is planned for OS 4.1
Progress report today from the head of Hyperion:
“Still plenty of work to go around but we’re getting there.”
“Me? I just hate it. The screenshots just look awful. I could build a Photoshop mockup of Amiga’s UI that look much more
professional and less, well, ugly. Maybe it is appealing to old Amiga
user, but..”
No, the look of those screenshots isn’t appealing to anyone, but they
are there to show what features will be added and how it can be
customised.
To see what a well set up Amiga GUI can look like, go here:
http://www.nowee.org/overshaker/index.html
The point is that those shots at overshaker were done using several
patches and hacks, and in OS 4 all those will be properly integrated
into the OS. I’m sure there will be plenty of themes to select from a
few weeks after OS4 hits the streets.
There was some initial confusion regarding whether the ExecSG kernel was coded in 68XX assempler or PPC compatible “C”.
This has now been unambiguosly answered by Thomas (The programmer responsible for the coding)
Quote:
I guess I have to clean up this confusion, even though I wanted to stay away…:
Believe me (as the main programmer behind ExecSG): ExecSG has *never* been 68k. It was being written on the PPC from day 1.
I think it wouldn’t make any sense to start such a PPC native OS as a 68k assembly project.
Unfortunately, the wording in the feature list was a bit misleading: It refers to the *OLD* exec, not to ExecSG.
Ok, I’m off again
Thomas
/end quote
If you want to, check out the link yourself on Amiga.org http://amiga.org/modules/news/article.php?storyid=864&comment_id=58…
Regards
Darren
I thought that they said in the beginning that BCPL was the answer to operating system implementation and now they’re using C…hmm.
To see what a well set up Amiga GUI can look like, go here:
They still look ugly. The icons are still the same. And so on. (Besides, I do hope the final release would see a comestic change).
If they don’t know how to do so, I could work for two months with them, November and December, make it look professional :p. (I know this would never happen..)
bousozoku, BCPL was an older language form which the C language was indirectly derived. Major parts of AmigaOS was written in this programming language.
Rajan R,
> If they don’t know how to do so, I could work for two
> months with them
They know how to do so themselves. They have some very talented graphics artists working for them. The provided functionality as being demonstrated in the screenshots are by far more important however, than any icons.
Actually, only portions of AmigaDOS were written in BCPL. Not the OS proper (the exec kernel). When AmigaDOS was re-written in C, many data structures still relied on BCPL pointers and strings which are different than C pointers and strings.
AmigaDOS was actually purchased from another company, Metacomco I believe, and adapted for the Amiga because the company just didn’t have the time to do it right.
None of the exec or GUI structures rely on BCPL at all.
Now, I thought that the original exec was written mostly in C and then re-written in 68k assembler. AmigaOS being an OS developed with C was supposedly one of its strong selling points.
The biggest part of the AmigaOS is written in C, the time-critical parts in assembler.
> Not the OS proper (the exec kernel).
What do you mean by “proper”? The Exec “micro”-kernel is only a very small part of the complete OS.
> AmigaDOS was actually purchased from another company,
> Metacomco I believe, and adapted for the Amiga because
> the company just didn’t have the time to do it right.
TripOS (MetaComCo) was originally planned for usage with the Amiga computer by Commodore. However the Amiga team was very unhappy with this OS as it severly lacked features. A portion of the TripOS code was used in the Dos.Library, eventually in v36 it was replaced by a combination of assembler & C code.
I find the screenshots remind me of a slightly more polished workbench circa 1995/6 (sorry I can’t be more specific about the version).
However, I have _never_ thought of the Amiga UI as aesthetically pleasing and AmigaOS4 doesn’t change my opinion on that. I think the UI is ugly.
>What do you mean by “proper”? The Exec “micro”-kernel is only a very small part of the complete OS.
Well, yeah. I guess that I wasn’t clear. I understand AmigaOS to consist of Exec (the kernel), Dos.Library (the part that contains some legacy BCPL structures) and Intuition. And of course all of the support programs that fill out the details.
I haven’t looked at my Amiga books in awhile, but I think that only Dos.Library contains any old BCPL based code (well, structures anyway). Any actual BCPL code was rewritten when AmigaOS 2.0 was released.
>TripOS (MetaComCo) was originally planned for usage with the Amiga computer by Commodore. However the Amiga team was very unhappy with this OS as it severly lacked features. A portion of the TripOS code was used in the Dos.Library, eventually in v36 it was replaced by a combination of assembler & C code.
Amiga Inc had started writing their own dos code but they ran out of time and they had to adapt the Metcomco code to the new computer. I agree that they re-wrote the code but weren’t they still stuck with such things as BSTRINGS and BTPRS? I haven’t read the story in some time though and my facts could be off.
I run OS/2 (eComStation.com) And let me tell you that Amiga OS 4 is pretty comparing to eCS. The UI enhancements are set for eCS 1.1
When the new Amiga Inc. (former Amino) has been founded
the goal was to create a revolutionary new computer.
The first computer on the world, that would run completly
in software !
This is why they have chosen TAO’s elate/vp system.
A VP is a “virtual processor”. That is, the CPU is completly
done in software.
The vision is very interesting: With complete hardware-
abstraction it would be possible to have your PC on a ZIP
disk, CDRW or CompactFlash, put it into any supported hardware,
be it a PPC, an x86, a MIPS, ARM or whatever architecture
and boot. There would have been two systems, a hosted one,
running on another OS, as a program, versions exist for Windows
and Linux, and a native one, that would overtake the hardware
below. Since most consumer-hardware is x86 anyway, noone really
saw the vision. But in the future, and Amiga Inc. were well
aware of that, we will see (it actually started already) a merge
of consumer electronics (DVD player, VCR, Radio, Amplifier,
TV, CDDA player etc.) and computers. The homes will be computerized.
Maybe not in 5 years, but this might start then. All we need
is cheap (and I mean it) flatpanel TFT/LCDs with good video
cabability and touchscreen support. Big industry is planning
the Firewire connected home anyway, also a project, Amiga Inc.
is well aware of (www.havi.org)
But TAO refused to add the extensions to their barbone OS,
that Amiga Inc. would have needed.
Now the situation was typical for Amiga Inc.:
First they abused the community for their own way, that no-one
in the community (except some nerds) wanted to follow. All
wanted a new Amiga, being just as the old one, but more powerful.
And they wanted their beloved AmigaOS.
Both things Amiga Inc. refused to offer. Still, they announced
their products by being supported by “the programmers of the
Amiga community” (or similar words).
Then they stepped back. Big steps. They started planning
the AmigaONE, even later they decided for AmigaOS4, just
after it got clear, they’d need a RealOS(tm),.since TAO’s
engine was good for small devices, but not servers or desktops.
At the moment the plan is this:
Have a hardware independant AmigaOS4. At time of this writing,
the OS got a HAL (hardware abstraction layer) and got ported to C.
It still is very much the AmigaOS we know, but with major and
much needed enhancements, a big step forward for all of us
Amiga freaks. (And maybe for others as well).
This AmigaOS is tied to licensed PPC platforms. You can’t
buy a MAC and install it. Apple would need to license a
special dongle, that Amiga Inc. requires all AmigaOS-PPC
vendors to mount on their boards, in hardware.
So much about hardware independancy 😉 (pun intended)
But AmigaOS4 is not the last step.
AmigaOS5 is planned to merge the existing OS technologies
from both sides and see the world’s light as something
completly new: A brand new OS. Suitable for desktop, embedded
(PDA, cellphones etc.) and server market.
The whole system can scale down. You will be able to buy
Quake-VIII and have *the* *same* binaries run on your cell-phone,
desktop and if you want, server, as long it has been ported
to AmigaOS and the machine is in fact running this new
AmigaOS.
For the consumer this is a dream come true. For the software-
producer it might be so as well:
no porting efforts = more resources on cleaning out bugs,
one API to learn = sell the software on many different markets
and – less advertising (at least in theory), but this would
heavily depend on how much AmigaOS5 will be advertised. Something
that needs to be seen…Amiga Inc. is *very* low on both resources
and capital, I’d say. On the other hand, Amiga Inc.’s strategy is
interesting. They like to talk about themselfes as ‘content-delivery’
company. And in fact, this makes it easy for them to have
alliances with major software houses. In fact, a software-
vendor does not need to advertise its product on different
platforms either. It is enough if people know: It was written
for AmigaOS. And we all know (hopefully) that if it has
been written for AmigaOS it will run “anywhere”. So it will
be interesting for these software houeses to advertise, to
promote AmigaOS as well.
The real problem I see at the moment is that Amiga Inc. is
too small minded. They also loose one interesting partner
after the other. And they are too egoistic, concerning the
community. They could have done much more clever PR, but
they failed. And – the hardware-market the new Amiga
(you know, the one, that runs in hardware) is targeted at,
is not yet really existant.
We’ll have to see. It can fail easily.
As for the question why they have chosen PPC: This is quite
easy to answer. AmigaOS development stopped largely in 1991.
The market was different at that time. It was either x86
or 68k. At least for the home users. Then Amiga users wanted
more speed and around 1994 the first accelerator boards showed up,
having a PPC, which would take over computing intensive tasks
from the 68k, that still ran in parallel. The PPC was a co-CPU,
making these expanded Amigas SMP systems, actually.
This was at unstable times, when the Commodore IP got sold
to many companies and the hick-hack was going on, which CPU
to use, if at all (coz’ noone really wanted a new Amiga, all
wanted the IP rights and patents or brand). Commodore did
some new architecture based on MIPS, if I remember correctly,
but this got never finished. Then someone announced PPC.
Then this got refused. Then again and so on.
Now the PPC did have some connection to the Amiga, this is
why they finally have chosen it, also being more popular
than other exotic CPUs.
It is ironic, however, to see the PPC being a close-to-dead
platform, now that AmigaOS finally gets finished and we
get a new Amiga, that in fact, is a PC with a PPC instead
of x86.
Sorry, I made a bad typo:
[…community. They could have done much more clever PR, but
they failed. And – the hardware-market the new Amiga
(you know, the one, that runs in !!! hardware !!!) is targeted at,
is not yet really existant…]
‘!!! hardware !!!’ should be ‘software’
‘scuse me.
I run OS/2 (eComStation.com) And let me tell you that Amiga OS 4 is pretty comparing to eCS. The UI enhancements are set for eCS 1.1
True, true, eCS 1.0 looks like something from the late 80s.
One of my former bosses was in charge of accounting at Commodore. He seemed to have a string of companies he ruined financially prior to Commodore and where I was working, which required bankruptcy protection the first day I started to work. “Welcome to the company, we have filed for chapter 11 protection.” Talk about a wonderful beginning.
It seems to me, though, that AmigaOS was originally an advanced class project at Cambrige University. I’ve got the BYTE magazines around here. If I find the articles (they’re somewhere within 1984-1986) I’ll let you know.
I definitely remember the MetaComCo advertisements for BCPL for both Amiga and Atari ST machines.