Hi alx,
your reply is a reasonable one,
BTW alx and The_Editor dont take my replies personally,
I am responding to your statements rather than to you as
people!
:my objection being that I used to also believe all the
hype about Hyperion coming out of
www.amiga.com,
I even started enthusiastic threads on
www.amiga.com about
Hyperions OS4 lists ("now this is good news" or something)
but over the months the second hand info from people
supposedly close
to the OS4 process has proven to be totally untrue,
as such I wont let people pull the same trick again,
Before I reply to alx's post, some of the posts in this
forum have answered some unasked questions for me:
it would appear that Gateway once they acquired Amiga
ransacked it for patents and IP, what they have left
behind is a gutted product.
And all that bollox about IP theft, Gateway are the
real IP-raiders, Bill is a small fish in all this.
I think Eyetechs refusal to use classic h/w in the A1 is
actually because the classic h/w IP has been raided,
One other thing, regarding Mr. Big, ISTR AmigActive magazine
mentioning some investor called
---------The Bowman Group-------------
(my memory is a bit vague, I may have recalled the name
wrongly), anyone know anything about this??
=========================================
Quote:
who is Rogue? One of Hyperion?
Correct
=========================================
Can you specify which one?
How many people at Hyperion are working on OS4??
============================================
Quote:
If so then I think you are indulging in wishful thinking
================================================
ExecSG, the kernel of OS4, has been seen by many people to run on an A1.
===============================================
this is the bit that I doubt, this is second hand info about OS4,
In January this year similar second hand info said OS4 is almost complete,
I was led to believe it would be perhaps a fortnight, then later Bill
announced the official release would be some computer show (C-bit or the
other one?). It didnt happen.
How on earth can you know how many people have *really* seen it?
Were you in a room with OS4 on A1 and saw eg 10 people looking at OS4,
or are you doing a so-and-so knows someone who knows the Friedens
type of head count.
Have you seen it yourself?
or are you a 68k betatester? (which would prove nothing)
There is so much disinformation about what has been achieved,
BTW I shall be cataloguing the URL to your above posting and any other
OS4 claims, so that in future I can parade the alleged claims and compare
this to the reality.
IMO until ExecSG is shown to the public it doesnt exist.
I realize that Hyperion probably dont have much cash, but they can
do their own small show in their own town or something.
=======================================================
It is the core of the OS, and contains (in the HAL) virtually all of the
platform dependent code. That was a while ago, and it wouldn't surprise me if a
lot more is working on it now.
=========================================
if you believe the PR then yes,
for ages they have talked about the HAL, they also had endless OS4
lists on the
www.amiga.com web pages, but its no use showing us
web pages of lists. Look, I can cut and paste their OS4 lists and
say I have done that. Look I have here on my desk a cure for cancer
and a cure for AIDS and over here I have my anti-gravity abstraction layer
etc. I can say anything, its the internet!
Show the public the actual stuff,
stop talking + listing and start showing, this would create a
big boost in faith in their project.
==============================================
But seeing how they've been shot down in flames before,
Rogue is currently not visiting Amiga forums.
===============================================
fair enough, but if they showed us even OS3.0 reimplemented on
PPC they would be given the red carpet.
If they reimplement OS3.0 on A1, no other stuff, I would be
queuing up to buy it. The other stuff can come later in an
ongoing incremental way.
===========================================
I don't know what The_Editor's sources are,
but I wouldn't be surprised if it's true.
=========================================
as true as your claims ie at least 3rd hand info
from my POV.
Perhaps you are both right, but I am very sceptical
compare Eyetech who talked the talk and then
not too much later walked the walk.
================================================
Quote:
if Hyperion had OS4 booting on an A1 they would firstly not
be pussy footing with getting OS4 on the god forsaken PPC cards, (burn another 6 months),
>Do you know much about the process of migrating AOS to PPC?
I know that it is a waste of time,
The correct procedure would be to cross compile OS4 onto A1
via eg a 68060 machine.
The FIRST thing to cross compile is the boot,
the boot up sets up the memory system, if you wish to have
clever memory functionality eg
VM (hard disk transparently extending RAM) and task protection
(progs cannot trash other progs, so progs crash alone) then
you have a lot of work to do. The boot up also has to
create out of a vacuum all the Exec infrastructure,
task scheduling, library management, etc
the boot up also has to get the file system up and running,
so that s:startup-sequence can get executed,
By the time the Workbench screen appears all the really difficult
stuff has been done. The OS infrastructure is an up and
running thing: memory infrastructure including VM and task protection,
the library management infrastructure, the filesystem,
from here onwards the problems are all traditional programming problems.
This is why I was insisting on seeing the boot,
it is Step no.1 in creating an OS and it is probably the most
difficult problem in Computer science.
As you know Berndt Mayer has sidestepped this self referential problem
by hacking the Linux kernel,
Pegasos sidesteps it by Morphos being a subsystem of the quark kernel,
hence Morphos inherits all the functionality of quark which seems
quite advanced.
These guys are sidestepping it because it is so difficult,
note that even Apple I think uses some external kernel,
Classic Amiga did their own kernel, by Carl Sassenrath,
the power of AmigaOS all comes from the power of Sassenraths
boot!
The original AmigaOS was cross compiled onto the A1000 from
AppleMac using Lattice C compiler (this is how I understood things).
=========================================================
>The PPC equipped "Classic" Amigas still have custom chips etc,
the fact they have the custom chips ironically is the whole problem!
The tricky thing about A1 is it *doesn't* have the custom chips,
if you have the custom chips to help you then it is easy-peasy
(in a technical sense, could be months and months of work though)
Its cheating to port OS4 to PPC, firstly as you say it has the
custom chips, secondly it has a 68060,
They are procrastinating from the nightmare problem which is
OS4 on A1.
I have no doubts they can achieve OS4 on PPC, my doubts are
about OS4 on A1. Perhaps if they use a ready made kernel
which isnt Linux they may have a chance. Trying to go via
Linux is going to be really difficult as
Linux and Unix are bloated and impenetrable and incomprehensible
systems. Dont do it! I realise Berndt has done it though,
he joins the elite by having succeeded.
=================================================
therefore any port to PPC will be finished first on them.
============================================
it will be much easier and quite irrelevant to OS4 on A1,
The easiest will be OS4 on 68k which is more sensible
than OS4 on PPC,
=============================================
If Hyperion were to release the A1 version first,
then they'd have to hold back
ack a complese CSPPC version for a while. Great.
==================================================
Forget CSPPC! A G3 Pegasos II costs £299,
people with CSPPC should buy a G3 machine,
====================================
Quote:
I saw OS4 on an A4000 at WOASE last autumn, a tour of this is no news
and its not OS4 on an A1,
What you saw was OS4 modules on 68k AmigaOS.
=================================
So it was less than I was led to believe,
you've answered one thing I have been wondering which is
was what I saw PPC OS4 or 68k OS4,
you are saying it is 68k OS4.
Why not release 68k OS4?
===========================================
The OS4 tour had OS4 modules on ExecSG - a very nearly complete OS4.
=========================================
No! It was OS4 on PPC ExecSG, 10 times as easy as A1 ExecSG,
OS4 on PPC is a totally different problem from OS4 on A1,
they may have fooled you and Bill but they dont fool me,
Now if they bring the Classic Hardware to A1 then they can
reuse the PPC ExecSG, and it becomes relevant and could be used on A1.
Then it would be a true Amiga One,
But Eyetech say no way will A1 use the custom chips,
From a h/w engineering POV it must be straightforward,
And probably someone else owns the h/w IP so legally Eyetech
maybe cant do that. I dont know.
================================================
Quote:
when developers create even a prototype be it a car or a computer or an
OS they will parade it at every chance, they will shout it from the rooftops,
I dont see that, because: its not happened!
If they did that with everything, certain people would:
>1) Get fed up
Bollox! People would flock to OS4 and A1, you couldnt be more wrong!
The only people who would be fed up would be their competitors and
their zealots,
>2) Claim that they're not making enough progress based on what they've done before
This is so pathetic!
You are admitting defeat before the battle has happened and deciding
therefore not to battle.
Get a grip, you are WRONG!!
You have to take risks.
No pain, no gain,
======================================================
>Hyperion are a small company - they don't have the money for serious PR.
=====================================================
A small company manufacturing big excuses and big rumours,
it doesnt need money, just put an A1 + OS4 in 2 suitcases, hop on a train
through the english channel, emerge at Waterloo station, go on the
London underground (Northern line and Circle line I think) to Liverpool Street
station, and its a short train journey to WOASE,
Buy your long distance train journeys long in advance and they could be
considerably cheaper.
================================
Quote:
remember how Eyetech would show off the A1 at every chance,
long before the Earlybird, they even had photos of their PCBs on their
website back in 2001 (I think),
>And now people can look back at those, and wonder
>"Why did they waste time on those designs?"
No! no! no!
It shows they arent afraid of risk, they put their necks out,
they made some miscalculations: this *always* happens,
they then redesigned and delivered last summer.
To hit the big time you have to risk total failure,
winning is about risk,
You are proving again and again that OS4 is a very timid effort,
=========================
Quote:
Ben has vanished from the scene he used to be ever present on ann.lu,
>Simply because of all the pressure and insults he was getting there.
people are fed up of all talk and no action, (by talk I include the
OS lists),
As I say if they merely reimplement OS3.0 into G3 native on A1 or
even Pegasos they will become very popular.
>Some people from Hyperion have even abandoned Amigaworld.net
(generally a very peaceful forum).
why?
lack of credibility for not showing us stuff???
Has Mike Bouma excommunicated Hyperion??
If Mike is against them then its the beginning of the end,
The last bastion of Amiga Inc has crumbled
==================================
Quote:
why dont they learn from Hack and Patch that PPC cards are a lose-lose thing
>Tell that to all the Amigans who spent loads of money on them.
===========================================
I just have! (to all the PPC owners reading this),
Advice: dont buy a multi CPU machine until it is proven beyond
a shadow of doubt that the OS is sound on them.
IMO multi-CPUs are a stupid move, its far more generic and
extendible to network together single-CPU machines. Its straightforward
to network say a 1000 single CPU machines, try OTOH making 1 machine
with 1000 CPUs, remember transputers?? you could link 4? of them
together, it never really took off. I knew a guy who was studying
computer h/w at university and he connected a transputer to his A1000,
he had worked at the transputer company as a holiday job.
The problem with the multi-CPU approach is that it requires a lot
of very specific work to do it both on the h/w side and OS side,
by the time the work is complete the CPU will probably be obsolete.
There are very tricky unforeseen OS bugs that can emerge on a multi CPU
setup, avoid the concept.
================================
Quote:
with OS4 the only original work I see is some tweaks to intuition and workbench,
the rest is all ports and 3rd party work eg AmiDock,
:just recycling other peoples work, if OS4 ever comes out it is going to be
such an anti climax, still its based on the original source code
so it must be wonderful.
?
Perhaps you've been mislead - I'll explain;
>Hyperion have re-written the AOS kernel
so they claim, perhaps for PPC yes,
they are too weak and fearful to show it to us if it does exist,
>They have created an integrated emulation solution which many believed impossible
>Whole modules like AmigaDOS have been completely re-implemented in portable C
>No OS3.5/3.9 source code was avalible - Hyperion had to go from 3.1.
You said "whole modules" and not "all modules",
apart from AmigaDOS and Exec are there any others??
Why not just give us 3.0 or 3.1 for A1? Forget about 3.5 and 3.9,
Start afresh on the A1 from 3.1,
you have to make compromises in real life,
JIT emulation is 75% of native, from a theoretical POV this means
its as good as native.
Why not just use JIT emulation for the OS3.5 and OS3.9 CD's?
==============================
With a handfull of people and meagre rescources,
the work they have done is amazing.
===================================
More amazing than the work of 2 people Ralph and Berndt?
amazing vapour yes,
==============================================
OS4 is not "just" a recompile from PPC -
it's as big a difference from 3.x as that was from 1.x.
==============================================
How do you know this?
Used it yourself?
wishful thinking?
dont make claims on something you havent seen,
AAA was a wonderful thing, but it never happened,
effectively its as vaporous as OS4 on A1,
==========================================
Quote:
The fact that after 2 years a games company cannot show us OS4 on A1,
proves to me that OS's are much more difficult than games,
the OS is not a game, GAME OVER Hyperion, 0 points,
>Exactly - do you think they expected to manage it quickly?
if they hadnt got OS4 booting on an A1 and shown this to the public
within lets say 6 months of starting they should have done the
honorable thing and given up.
BTW Hyperions AmigaOS source code license, what is the small print?
1. Is it an indefinite license?
2. Is it transferrable?
3. Can it be revoked by both parties?
4. Is it exclusive? (could eg Berndt apply for a parallel license of it)
5. How did they get it?
==============================
Quote:
The arrogance from Amiga central is so sad,
Berndt and Ralph are years ahead of them, so is UAE, cloanto,
3rd party cybergraphics.library pisses on graphics.library,
etc there is a whole roomful of people ahead of the official Amiga,
>Cloanto and UAE ahead of OS4? You make me laugh
Fair enough! But they exist OS4 on A1 doesnt so in that sense
they are ahead which was my point!
Merely by existing the Spectrum computer is ahead of OS4 on A1,
A line of beans is ahead of OS4 on A1 because it exists,
>Cloanto make a UAE bundle.
>UAE is an emulator, not an OS.
I agree, emulator notequals OS,
but it exists,
>Berndt's Amithlon is no way as complex as OS4,
according to the second hand info,
anyway complexity is irrelevant,
badly written things can be exceedingly complex,
I want simplicity and functionality, not complexity.
If you want complexity get L-unix. (horrible things)
>as it's still based largely on the 68k code.
so what if its emulated at 75% speed, just as good as native,
>And what relevence does CGX have?
its a 3rd party effort, my point was that the 3rd party is ahead
of AmigaOS,
>OS4 uses P96, which is 100% compatible.
The fact that it uses P96 is AmigaOS admitting defeat,
Quote:
my A1200 would be moth-balled if it werent for a 3rd party 68030 accelerator,
>So why were you earlier complaining about the use of 3rd party stuff?
Not sure what you are referring to, is this a reference to PPC?
I objected not because its 3rd party but that it is irrelevant to the
OS4 on A1 concept amongst other things.
Development and testing should be directly on an A1,
cross compilation should be done from a 68060 machine onto the A1,
I would expect the RAM to be used as a ROM as was the case for
kickstart on A1000.
The boot is a tricky chicken and egg problem, this trickiness is at
the heart of the whole issue.
Maybe you can create a disk on the A4000, connect this disk to
the A1 and this disk initialises the A1, once you have enough functionality
you can then develop directly on the A1. Creating a new OS is like a virus,
the virus has to transmit itself from an existing host into a new host,
its difficult. The masters of this process are Unix, the c language was
invented by Dennis Ritchie in order to port Unix to new alien hardware.
Part of the unix strategy is that their c compiler cross compiles itself
onto the new hardware like a virus moving to a new host.
Unix is an elaborate virus, L-unix (Linux) is a freeware aping of Unix.
Anyway PPC shouldnt be entering the equation.
=========================
The danger is that when OS4 finally gets released
The G3, G4 ,.. series (Gx?) will be discontinued,
