@pixie

Quote:

pixie wrote: @Hans Quote: I've never used Amithlon so I have no idea how it compares in functionality to Amiga OS4. I think it's safe to say that the amount of work required was much less. It's based on a modified Linux kernel, did not come with a USB stack, and, the JIT emulator was based on previous work. I can't seem to find out if the exec was rewritten for the x86, or runs under JIT. I'm guessing that it runs under JIT, i.e., no rewrite necessary. No, the work was underneath, and it doesn't mean it was simpler, only the approaches taken were different. /me wonders if 68k software running on a 3Ghz Amtihlon pc doesn't run faster then PPC on A1...

What do you mean, "the work was underneath"? Amithlon is really an emulator that allows Amiga OS 3.x to run on an x86 box. It's not a native implementation of Amiga OS. Looking at the work required for Amithlon, I think it would take less work than Amiga OS4. It's essentially an emulator running on a Linux kernel, without chipset emulation, but with some drivers for standard hardware. That's less work.

Note, I'm not saying that Amithlon is a bad product. I wish that it were still available. However, it's not an OS update. It's not just a different approach; it's a different kind of product.

Hans

_________________
Join the Kea Campus - upgrade your skills; support my work; enjoy the Amiga corner.
https://keasigmadelta.com/ - see more of my work

@Hans

Quote:

Looking at the work required for Amithlon, I think it would take less work than Amiga OS4.

That is probably true. Of course, it took only one man-year, too.

Quote:

It's essentially an emulator running on a Linux kernel, without chipset emulation, but with some drivers for standard hardware.

Actually, it's considerably more than that.

See, just like the AmigaONE, your average PC has no CIAs. It has no custom chips. It has A1200 IDE device. Your average PC is essentially just an AmigaONE, with a different CPU, fewer broken-by-design components, and a heck of a lot more variety when it comes to disk controllers, graphics cards and network interfaces.

Yet poor lil' me did *not* have any AmigaOS sources. I did *not* have the option of changing anything --- AmigaOS needed to run unmodified. So every piece of hardware it expects needed to be emulated, at least to the level that is required for proper functioning of AmigaOS.
Yet at the same time, *SOMEONE* had to handle the actual hardware.

For disk controllers, that "someone" was Linux, meaning that I "only" needed to write a SCSI device replacement (and ever since that day, I am utterly convinced that whoever designed SCSI device in the first place should go straight back to uni and learn about software design. Yuck! What a piece of non-modular crap!). Of course, that replacement needed to translate between Linux's way of looking at disks, and AmigaOS's. I suspect that's actually a lot more work than adapting the A1200 IDE driver to a different IDE controller.
Then there are the CIAs, handling timing and input devices. Well, Linux PCs (in 2001) did not have high precision timing. So I had to implement some awful code to reprogram the periodic timer interrupt on the fly to actually get a decent source of high precision time. And of course I also had to write code to handle PS/2 devices (which, in the case of mouses, were many and varied in their protocols back then). And once I have all this information, I can't just feed it straight into input.device and timer.device, because, well, I don't have the sources, so obviously I can't *change" the sources. So instead of handing over the info I have to the OS, I have to convert it into values in emulated registers of lng obsolete hardware --- and in return, have to somehow extract what the OS wants me to do from what it writes into other emulated registers.
Graphics cards, same situation --- I get to write a P96 driver; I don't actually get to look at, let alone modify, the P96 system itself. So in addition to having to convert between Linux's view of graphics cards and P96's view of graphics cards, I get to create lots of difficult workarounds for things which would be a 30 second change *if* one had the sources. Oh, and of course using the linux drivers, while having advantages in one way, means that I have to deal with a large variety in capabilities available, all through one P96 driver. OS4 P96 drivers have a much easier time (and let's face it, if you actually have documentation for a card, writing a P96 driver is not a huge job).
And there are many more such areas, where the very fact that it is "only an emulator" means more work, not less. I suspect Hyperion found that out, too, hence their speedy change of course in 2002 :)

Of course, that's all just to get OS3.9 running. Then there is the whole x86 native interface, which is completely transparent to the OS and manages to put x86 execution threads on perfectly equal footing with 68k execution threads, being fully preemptively controlled by the Exec scheduler. Once again, this would have been *a lot* easier if I could have mucked around with the Exec sources (or had written my own Exec --- but Amithlon was not meant to replace AmigaOS, it was meant to run it, "It" being the real thing, the very same code people were running on the existing Amigas).


So yes, I agree that comparing the man-hours on Amithlon and OS4 probably does not provide a whole lot of insight. However, do not assume that Amithlon being "just an emulator" means it was any less work. At times, I would have given a lot for access to OS sources, and the permission to modify them.

@umisef

Thanks for the info. I wasn't suggesting that your work was not impressive. It's a real pity that we can't buy it. IMHO, not only were the design goals different (as you've confirmed), Amiga OS4 is a larger scale project. The scope of what they're doing is larger.

I actually wonder what's easier, emulating the CIA chips, or, getting rid of all the CIA chip dependencies. That depends on the Amiga OS 3.x source code. The source may well be riddled with CIA dependencies all over the place. I'm sure some stuff was easier for them, and other stuff was easier to support by emulating the original hardware.

Hans

_________________
Join the Kea Campus - upgrade your skills; support my work; enjoy the Amiga corner.
https://keasigmadelta.com/ - see more of my work

Quote:

No not at all but some were telling us these things (Ainc is insolvent) way back and we did not listen. I even remember the pizzas being sent to prove a point and even after the auction the "it's just unused office furniture". Alot of these facts are not new. Of course it is hard to figure out what is fact and what is revision of the truth, and I tend to hear what I want to hear and disregard the rest (like most people I guess) to my chagrin often. Besides I am easily amused.

Yeah, TBone and I both sent pizzas to the offices back in 2002, Ben and MikeB were among those telling us that didnt matter.
-Tig

_________________
We played the first thing that came to our heads, it just happened to be the best song in the world.

Quote:

kgrach wrote: @Tigger I have paid for every piece of software I own and all of my downloaded music too.

Ok so either you are a silly customer and didnt register my software, and I'd love to hear what you think is wrong with it, or you havent ever owned my software and you owe me an apology, which is it?

Quote:

Nobody else offered to do the port.

Not true at all, much like the bidding on 3.5, the lowest bid was taken, not the only offer.

Quote:

You either are ignorant of the Amiga scene or you are purposely distorting the truth.

Not ignorant at all, ask around, not distorting the truth either.

Look at what I said.
Quote:

Porting the OS to PPC should not have been a 5 man year effort, much less the 20+ man year effort this turned out to be. It was an OS primarily written in C that they had to move from 68K to PPC, thats not a 10,000 hour job much less the 40K plus hours they seem to have burned on it.

Quote:

The AmigaOS was written in Assembler C and BCPL. But you knew that so why imply it was a simple C port!

I said it was primarily written in C as you quoted above, it was primarily written in C. In fact more of it was written in C then before, because Olaf had cleaned up the code as Hyperion told us in there documentation.

Quote:

You are also neglecting to mention the Uboot port and HAL for the A1and micro not to mention the monkey wrench Allen threw in by using a new CPU at the end

UBoot and HAL are 1 month efforts, period., actually shouldnt take that longe. Green hills ports to entirely new boards in 2 weeks.

_________________
We played the first thing that came to our heads, it just happened to be the best song in the world.

Quote:

NutsAboutAmiga wrote: @Tigger Draco where a hardware clone it where deigned to be compatible, whit AmigaOS, Draco did even have the Zorro slots, this point me to yet an other module that need to be replaced on AmigaOS4 the expansion.Library, expention.library now support the PCI bus.

Draco doesnt have any of the custom chips except the CIAs, yet the OS ran on it fine, Draco isnt the only Amiga like clone that was created over the years, its just the one most people know about. Since all the libraries you mentioned run fine on it, implying the have to be completely rewritten because of the AGA chipset isnt correct.
-Tig

_________________
We played the first thing that came to our heads, it just happened to be the best song in the world.

Quote:

CodeSmith wrote: @Tigger Quote: BCPL as I recall. Its not that much different the C, in fact C was based on it to some degree. You're kidding, right? This is some BCPL code, from the Wikipedia page on BCPL: One can see a family resemblance, but you're not going to do a whole lot of cut'n'paste when you're porting something that looks like that to C.

Dont have to cut and paste, there a literally dozens of Open Source converters available, and I'm sorry by BCPL is one of the ancestors of C, thats pretty much agreed to by everyone. In addition, I'm not sure the Olaf hadnt already done the conversion when this entire thing started.
-Tig

_________________
We played the first thing that came to our heads, it just happened to be the best song in the world.

O.T.

Ok every one, it's been a long torturous thread and it's time for a commercial break.
Tig you might like this one....... G4TV - Code Monkeys

Plaz

Last edited by Plaz on 26-May-2007 at 05:02 AM.

@Hans

Looks are irrelevant, so are USB and TCP/IP as all these things have been done by 3rd-parties, and to some degree even started much earlier. So they can't be used as an argument why it tool so long to get OS4 up on it's feet.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@Kronos

As a programmer (and a not so good one, I admit this) have you ever tried to do a port or actually looked at the mess known as AmigaOS code?? I don't think considering resources that this was a long time at all, it was probably about right...

@Tigger

Heavens to BPTRs, where's Tim King when we need him

@umisef

Quote:

I am utterly convinced that whoever designed SCSI device in the first place should go straight back to uni and learn about software design. Yuck! What a piece of non-modular crap!).

The same holds true for the rest of the classic AmigaOS, as far as I've been told by developers, so you can imagine what the OS4 developers had to deal with (scsi.device being perhaps 1% of the whole AmigaOS).

@Tigger

Quote:

Dont have to cut and paste, there a literally dozens of Open Source converters available, and I'm sorry by BCPL is one of the ancestors of C, thats pretty much agreed to by everyone.

If I were interviewing someone at the software company I work for, and he told me that, I'd politely thank him for his time and show him the door. I'm starting to believe kgrach's allegations that you may not be such a hot shot coder after all.

@CodeSmith

BCPL strongly influensed B programming language. C is descendant of B.

_________________
Amiga Developer
Amiga 500, Efika, Mac Mini and PowerBook

@itix

Yes, and to help keep things straight:

Quote:

Kernighan and Ritchie (The C Programming Language, Second Edition, Prentice Hall, 1988, page 1) wrote: Many of the important ideas of C stem from the language BCPL, developed by Martin Richards. The influence of BCPL on C proceeded indirectly through the language B, which was written by Ken Thompson in 1970 for the first UNIX system on the DEC PDP-7.

Last edited by meega on 26-May-2007 at 11:14 AM.

_________________

@Tigger

Quote:

Porting the OS to PPC should not have been a 5 man year effort, much less the 20+ man year effort this turned out to be. It was an OS primarily written in C that they had to move from 68K to PPC, thats not a 10,000 hour job much less the 40K plus hours they seem to have burned on it.

If you've been following the development of OS4 over the past 6 years you should be aware that at certain points the core OS4 developers have had to work on other projects in order to keep money coming into Hyperion.

The ¤800,000+ didn't come from nowhere and as far as I know Hyperion isn't run using someone else's money. I don't know how you or anyone else could outside of Hyperion could make a calculation of how many hours were spent on OS4 between Nov 2001 and Dec 2006.

Quote:

CodeSmith wrote: @Tigger Quote: Dont have to cut and paste, there a literally dozens of Open Source converters available, and I'm sorry by BCPL is one of the ancestors of C, thats pretty much agreed to by everyone. If I were interviewing someone at the software company I work for, and he told me that, I'd politely thank him for his time and show him the door. I'm starting to believe kgrach's allegations that you may not be such a hot shot coder after all.

I'm not sure I'd want to work with someone who doesn't understand that BCPL is the ancestor of C, because Kernigan and Ritchie surely do.
-Tig

_________________
We played the first thing that came to our heads, it just happened to be the best song in the world.

@CodeSmith
Quote:

If I were interviewing someone at the software company I work for, and he told me that, I'd politely thank him for his time and show him the door.

Then you'd be turning down someone with more knowledge of computer history than yourself, which may or may not be beneficial to whatever your company does.

Added:
Reference the original introduction to "The C Programming Language", if you have that on your bookshelf.

Last edited by jorkany on 26-May-2007 at 02:29 PM.

_________________
Here for the whimpering end

@jorkany

Quote:

Then you'd be turning down someone with more knowledge of computer history than yourself,

I reckon it depends on what he meant by "and he told me that". If anyone volunteers that kind of information during an interview, apropos of nothing at all, that's a bit odd(1). If someone says it in a context where it makes sense ("Name some of the defining influences on the C language" :), it's a sign of broad knowledge.

(1) When we last interviewed, we had one candidate who related all sorts of fascinating facts about PDP11s. Which was great, and rather amusing, but really had absolutely nothing to do with anything we had ever said, or advertised for, or were ever likely to need. It was one of very few candidates who never even got to sit the coding test....

@Kronos

Quote:

Kronos wrote: @Hans Looks are irrelevant, so are USB and TCP/IP as all these things have been done by 3rd-parties, and to some degree even started much earlier. So they can't be used as an argument why it tool so long to get OS4 up on it's feet.

I didn't mean "it looks different" as just the GUI style. Find an OS4 machine and start browsing through the drawers on the system drive. Look at the prefs, utilities, etc. You might notice some new stuff and existing apps/utilities have new functionality. 3rd party or not, Tigger seems to be including all work done by everybody involved in his man-hours figure.

While you're at it, download the SDK and start comparing it to the old NDKs. That might give you a hint of all the changes that have been made.

BTW, what are people counting as the point at which OS4 was "up on it's feet"? In terms of an "initial port to PowerPC" with some extra functionality; that was done with the first prerelease we got early 2004. I'd say that it was "up on it's feet" back then. Everything since then has been continual improvements.

Hans

_________________
Join the Kea Campus - upgrade your skills; support my work; enjoy the Amiga corner.
https://keasigmadelta.com/ - see more of my work

Quote:

Rob wrote: If you've been following the development of OS4 over the past 6 years you should be aware that at certain points the core OS4 developers have had to work on other projects in order to keep money coming into Hyperion.

I've been following Amiga Development since late 85, so I am aware that at times they did side projects, however given the over 5 years the effort has taken and the claims by Hermans over the years that over 30 people were working on the OS, calling an average of 4 man months per month of effort for the last 5 years is really probably pretty low, and that gives us 20 man years of effort for the project, which is really way too many.

Quote:

The ¤800,000+ didn't come from nowhere and as far as I know Hyperion isn't run using someone else's money. I don't know how you or anyone else could outside of Hyperion could make a calculation of how many hours were spent on OS4 between Nov 2001 and Dec 2006.

They didnt pay that money, they owe people that money Rob. If for example they had paid the Friedens, then Hyperion would own the code, etc for Exec-SG and the buyback clause would include it. I dont think anyone has been paid in full for their part of the OS. Thats why Hyperion cant let the OS go, because all those people are going to want there money.
-Tig

_________________
We played the first thing that came to our heads, it just happened to be the best song in the world.