Was anyone else disappointed with the A1200?

[jotd]

What, you're calling Thomson TO7 and MO5 (that we had in France) crap?

Well you're right.

[Gorf]

Originally Posted by Bruce Abbott

To think that Commodore, a calculator company that made their mark in low-end home computers used to play video games, could compete against SGI with their high-end 3D graphics workstations, is ludicrous.

Silicon Graphics started small in the 80s ... and much later than Commodore.
They just took a more professional approach.

Of course the Amiga fan will now tell us that SGI were also incompetent - I mean they could have used their 3D tech to make world beating gaming consoles that undercut the PlayStation, right?

They did:
the "Reality" coprocessor in the N64 was designed by Silicon Graphics.


Later (1999) a large group of important engineers left SGI for Nvidia.
So the successor of SGI-technology can now be found in the XBOX, the PS3, and the Nintendo Switch.
(and in the PC of course)

Some of us didn't need rapid GUI development tools for an OS that was defunct by 1996.

macOS is defunct since almost 30 years now? Why did nobody tell Apple!?

(macOS IS essentially NextStep by just an other name)

[TEG]

I came back on the XOR patent discussion.

This page indicate that Commodore had defaulted on payments. It would mean a deal existed with CadTrack but they stopped paying. Would perhaps explain a bit more why a new deal was not conclude with the complainant.

The CD32 was released in Europe and in Canada, but Commodore was unable to sell it in the United States due to a legal injunction for non-payment of patent royalties to a company called Cad Track. This was an early example of a patent troll, as the patent was for using the simple Boolean XOR (exclusive OR) formula for displaying computer graphics. XOR is a fast way to overlap graphics non-destructively and was commonly used by GUIs to draw bounding boxes. Every other computer company simply paid Cad Track their royalties as a matter of course, but Commodore had defaulted on payments and had refused to pay the $10 million penalty ordered by a US judge. At this time Mehdi Ali was still taking home more than $2 million a year.

Would be appreciated if someone is connected with its Google account so can access the page and bring back the content of the legal injunction link.

[Promilus]

SGI's premature announcement of its migration from MIPS to Itanium and its abortive ventures into IA-32 architecture systems (the Visual Workstation line, the ex-Intergraph Zx10 range and the SGI 1000-series Linux servers) damaged SGI's credibility in the market.

Credibility. That's something which does matter... SGI failed generally because they wanted to sell overpriced products with their direct competitors offering basically the same functionality (yes, as PC evolved). Amiga was quite different. It did offer for quite some time BETTER A/V subsystem for LESS. But there was limited professional software base (which SGI had kind of embedded in their workstations).

After 68k IRIS series they went MIPS. And after a while bought MIPS to secure (rather) supply and/(than) development. Basically what Commodore did with MOS. But later on bet on ill-fated itanium. Was MIPS developed poorly? Or was there a problem with professional graphic accelerators ? Who knows. Workstation GPUs were specialized - not by hardware but rather through firmware and drivers. Quadro is basically the same chip as GeForce. Of course CPU performance was still very, very important and I guess at some point lack of progress in MIPS development held them back. But I don't really get it. AFAIK PPC workstations for rendering ended up being much more popular. The exact route H&P and Apple went back then - and at a time that was a good idea.

So... A1200 was released in late '92. SGI went MIPS in '92. Apple was exploring RISC since late 80s and made an alliance with IBM (And Motorola) in early 90s. And obviously Commodore did think about that as well with Hombre. Little late but... would PA-RISC be really something affordable? They went for it to be NT compatible

According to Hombre designer Dr. Ed Hepler, Commodore intended to produce an AGA Amiga upon a single chip to solve the backward compatibility issues. This single chip would include Motorola MC680x0 core, plus the AGA chipset. The chip could be integrated in Hombre based computers for backward compatibility with AGA software

And here you were laughing about the idea of C64 compatibility in Amiga through the use of minimal required set of original chips as an optional expansion. So - using your arguments - it's easier to just get A1200 and set it up beside new Hombre to use classic amiga software It did sound funny for you when presented by me and here you go similar idea by C= engineer :P

[Bruce Abbott]

Originally Posted by Promilus

Of course CPU performance was still very, very important and I guess at some point lack of progress in MIPS development held them back. But I don't really get it.

Same reason the Amiga lost, PCs were swamping the market and everyone else was being squeezed out. Once PC performance matched what they had it was all over for SGI.

So... A1200 was released in late '92. SGI went MIPS in '92. Apple was exploring RISC since late 80s and made an alliance with IBM (And Motorola) in early 90s. And obviously Commodore did think about that as well with Hombre. Little late but... would PA-RISC be really something affordable? They went for it to be NT compatible

This is why I keep saying Commodore needed to die when they did, before they completely screwed it up.

And here you were laughing about the idea of C64 compatibility in Amiga through the use of minimal required set of original chips as an optional expansion.

Er, no. For sufficient compatibility you would need practically a full C64 in there.

So what would you gain from an optional 'C64 compatibility' expansion? Existing C64 owners who bought an Amiga could just continue using their old machine for C64 stuff, like owners of other home computers did. Existing Amiga owners who for some peculiar reason decided they wanted to run C64 programs could just buy a C64 (which by that time were dirt cheap).

The C64 and Amiga were totally different platforms, only having the manufacturer's name in common. So there wasn't any compatibility to carry over. Fans understood this. They bought an Amiga because they didn't want to run C64 software.

So - using your arguments - it's easier to just get A1200 and set it up beside new Hombre to use classic amiga software It did sound funny for you when presented by me and here you go similar idea by C= engineer :P

This is a different situation. Presumably new Hombre-based Amigas would still have the same OS and a high level of compatibility for software that used it - basically everything except hardware banging games. Furthermore by this time Moore's Law was making it a lot easier to do such a thing. Today we can put an entire Amiga into a 'cheap' FPGA, and if done in ASIC it would be even cheaper. That was not the case with the C64 in 1984.

I do object to the idea though, because it meant continuing down the path of competing against PCs by using a fancy graphics chipset, which was bound to fail. Commodore eventually realized this and was looking at putting PCI slots into future models. They were also looking at going RISC because 68k was going away. Then there were problems with Amiga OS itself. At some point it would have to get a new kernel too, like Windows and Mac OS did - if it was to compete against PCs.

Whichever way you look at it, the Amiga as we knew it was coming to the end of the line. It would have been nice to get more A1200s and CD32s though, and perhaps other more advanced (but still 99.9% compatible) models that kept the classic design going a while longer.

[Promilus]

For sufficient compatibility you would need practically a full C64 in there

On what particular hardware aspect of C64 you make that thesis?

The C64 and Amiga were totally different platforms (...)So there wasn't any compatibility to carry over

Like Apple II and Macintosh, oh wait!

This is a different situation

Like Yoda said "No different. Only different in your mind".

Presumably new Hombre-based Amigas would still have the same OS

Amiga OS is not Windows CE or Android to be essentially the same OS regardless of CPU architecture. And Hombre was very, very far away from 68k. So you'd - at most - get AOS4 level of compatibility. Probably less.

That was not the case with the C64 in 1984

Essentials are basically only VICII and 6502 with some glue logic (which obviously doesn't cost nearly as much as whole c64). Anything else can more or less be emulated by existing Amiga hardware (and emulation software).

Commodore eventually realized this and was looking at putting PCI slots into future models

Yep, after years of building up Zorro ecosystem... and ofc failing with ZIII target performance.

the Amiga as we knew it was coming to the end of the line

It was bound to happen sooner or later but not because of Commodore, because of Motorola (and Motorola always provide details for their products eol date to clients). But that didn't mean Amiga death was inevitable. Only 68k Amiga line. Most manufacturers invested in RISC because it did provide pretty decent performance boost due to superscalar pipelined architecture which was barely under development for CISC at that time. As it turned out when CISC eventually overcame all issues with implementing those features, RISC processors barely retained performance crown and soon enough improved x86 attacked market after market reaching out even for HPC machines and dominating them. And no, that's not because of IBM PC compatibility. Not for servers and HPC, nor gaming consoles, industrial computers etc. But because those were processors fairly cheap, fairly powerful, produced in sufficient quantities by at least 2 main companies (so not single source like ColdFire) and greatly supported by oem.

It would have been nice to get more A1200s and CD32s though, and perhaps other more advanced (but still 99.9% compatible) models that kept the classic design going a while longer.

Well, sure.

[desiv]

Originally Posted by Promilus

Like Apple II and Macintosh, oh wait!

The Apple II card for the early Macs was basically an entire Apple //e on a card.
As far as I know, there wasn't any other good way of running Apple II software on a Mac until software emulators came along and required pretty fast MACs (i.e. PowerPC).
I am not aware of any for the 68k based Macs.

[Promilus]

@desiv
Well let's see then. LSI chip handled everything except graphics, ram and cpu itself. RAM and CPU were on board, graphics was handled by software subroutines running on Mac hardware. There was floppy controller on board. IIe card was able to access some peripherals from host machine and part of Mac RAM as well.

Now... C64 doesn't have floppy controller - it uses serial communication with DOS on drive itself (which means floppy controller is OUTSIDE). Serial communication - that one Amiga can handle directly and accurately. Now then, what else? Joystick? Yeah, it can pass host machine to emulation software. Keyboard the same. SID? Amiga can emulate it on Paula. Results aren't that good but it still counts. So what we are left with?
Well we're left with... 6510, VIC II, PLA and RAM. Or 6502, some PLD as glue logic and software emulation of 6510+pla bank switching. ROM can be uploaded as image in amiga RAM. You might even dump VIC II and emulate it's functionality on host hardware at the cost of lower compatibility. Of course it was fairly possible to put most of C64 internals in a single chip like that LSI for IIe card. It was relatively simple logic (in comparison to e.g. Agnus).

[desiv]

So it was almost all done on the //e card and not by the Mac.
Which is what I was saying.

[Promilus]

Which is what I was saying

And all I was saying is NO IT DOESN'T HAVE TO BE WHOLE C64! I did point out which functionality CAN be successfully passed to Amiga hardware without breaking compatibility. And as for IIe cards... LC series were already quite expensive and IIe card (which indeed boosted sales of LC machines and according to apple sources half of those had it installed!) also was quite expensive. So how bad would it be to get C64 compatibility card for Amiga 1000 well under 100$? Take a guess...

[Thomas Richter]

Originally Posted by Promilus

Yep, after years of building up Zorro ecosystem... and ofc failing with ZIII target performance.

Which would have been a good step. PCI at that time could already deliver what Zorro had, and what ISA did not have. PCI finally had a mechanism that is close enough to "Autoconf", there was no longer a need to "set the interrupts" with DIP switches, which you had before. PCI was an industry standard, so lots of expansions would have become available. Just "too late", as always.

Originally Posted by Promilus

It was bound to happen sooner or later but not because of Commodore, because of Motorola (and Motorola always provide details for their products eol date to clients).

Hardly. CBM was doomed before the move to Risc, actually, AGA was already "too little, too late". The entire business approach CBM was successful with was already no longer working when Amiga came into their hands.

Originally Posted by Promilus

As it turned out when CISC eventually overcame all issues with implementing those features, RISC processors barely retained performance crown and soon enough improved x86 attacked market after market reaching out even for HPC machines and dominating them. And no, that's not because of IBM PC compatibility. Not for servers and HPC, nor gaming consoles, industrial computers etc. But because those were processors fairly cheap, fairly powerful, produced in sufficient quantities by at least 2 main companies (so not single source like ColdFire) and greatly supported by oem.

But they were fairly cheap *because* of PC compatibility. The x86 architecture is actually quite ill-designed, and it only became powerful because PCs sold so well, and enormous amounts of money went into the pockets of intel and AMD. It took a lot of engineering to bring this architecture up to speed - money Motorola and IBM no longer had to develip RISC any further, and money that goes into the arm platform (from multiple sides) nowadays because everything is mobile today.

[chb]

Originally Posted by Promilus

@desiv
Now... C64 doesn't have floppy controller - it uses serial communication with DOS on drive itself (which means floppy controller is OUTSIDE). Serial communication - that one Amiga can handle directly and accurately. Now then, what else? Joystick? Yeah, it can pass host machine to emulation software. Keyboard the same. SID? Amiga can emulate it on Paula. Results aren't that good but it still counts. So what we are left with?
Well we're left with... 6510, VIC II, PLA and RAM. Or 6502, some PLD as glue logic and software emulation of 6510+pla bank switching. ROM can be uploaded as image in amiga RAM. You might even dump VIC II and emulate it's functionality on host hardware at the cost of lower compatibility. Of course it was fairly possible to put most of C64 internals in a single chip like that LSI for IIe card. It was relatively simple logic (in comparison to e.g. Agnus).

I think you are massively underestimating the neccessary effort and overestimating the benefits. How do you think that integration would work? The VIC II does not have RGB output, it even does not have YUV output, but luminance and a color signal (3.58 MHz carrier QA-modulated with the I and Q color components in the NTSC case). So either a separate video output or de-modulation to RGB and mixing/switching with the Amiga RGB signal. Or some fancy digitizer (in 1985?), or a redesign of the VIC. Oh, and software-assisted VIC emulation is probably ok for productivity applications, but then why bothering to have a 65xx on board at all instead of going full software emulation? And why should someone buy an emulator card when it does not emulate games well, and isn't much more usefull than a pure software approach for applications that weren`t the C64's strength anyway?

The IIe-card had a clear usecase (custom applications for the Apple II, mainly in the educational sector) and a user base that was willing to pay for that (educational institutions). BTW, it was half of the LCs sold to schools that had the IIe card, I guess for other customers the share was far, far smaller.
The C64 and its derivatives had only very moderate success in the education sector, so they could not rely on that market.

So the IIe card was released much later (1991) than your proposed C64 card (at the Amiga's launch?), could use much advanced technology to emulate a much simpler system (in terms of graphics/sound hardware), and had a clearly defined and financially strong target market. I therefore do not think its success supports your arguments.

[Promilus]

@Thomas Richter

PCI was an industry standard

Yes but it came as such only in Pentium era. Which - on larger scale - is rather second half of 90s.

CBM was doomed before the move to Risc, actually, AGA was already "too little, too late".

In that particular phrase I was talking specifically about 68k-based Amiga (Amiga as we know it - by Bruce). So I stand by my opinion about it. It was bound to happen regardless of what business decisions Commodore made just because Motorola was killing that line.

But they were fairly cheap *because* of PC compatibility

And what exactly do you mean by that? Servers had different form factor, PSU, sometimes even lacked graphics, sound and ps/2 ports for input devices. What exactly from "IBM PC" world made it successful?

The x86 architecture is actually quite ill-designed, and it only became powerful because PCs sold so well

Well show me platforms with same expansion features, which you can assemble yourself and buy all necessary components in computer store around the corner. Exactly. It was popular not because of "IBM PC compatibility" which was long out of date and irrelevant. It became extremely popular because it was fairly cheap and most innovations there came out through collaboration between multiple vendors and adopting common standard. It had nothing to do with original IBM PC compatibility as most of that particular platform tech was obsolete by that time. And not entirely perfect compatibility was retained MOSTLY by CPU compatibility itself.

@chb

I think you are massively underestimating the neccessary effort and overestimating the benefits.

and vice versa. That's all theoretical speculation. My side, your side...
One thing is clear. A1000 came out in 1985. With little amount of software. And rather without spectacular financial results for Commodore. C128 came out in 1985. Was overall bad computer. Most innovations were implemented stupid way (faster CPU with good old VIC II? VDC chip? Faster drive which has to revert to 1541 mode for some games? nearly obsolete Z80?). But it was C128 which did sell well and produced results. Deep down you already know it wasn't healthy situation for Amiga propagation.

but then why bothering to have a 65xx on board at all instead of going full software emulation?

Because it takes 030 to emulate it purely in software at decent speed and with decent compatibility.

So the IIe card was released much later (1991) than your proposed C64 card (at the Amiga's launch?), could use much advanced technology to emulate a much simpler system

And yet it seems nobody here is able to answer the question why the hell it would've been necessary to put CIAs and SID and ROMs up there... And that's a half of C64 chipset. So... any particular idea why - if it was indeed going to happen - it would require all those chips? And if it wouldn't and could be made rather cheap - why would it failed when paired with over 1000$ machine without much of the software?

[TEG]

Originally Posted by Thomas Richter

Which would have been a good step. PCI at that time could already deliver what Zorro had, and what ISA did not have. PCI finally had a mechanism that is close enough to "Autoconf", there was no longer a need to "set the interrupts" with DIP switches, which you had before. PCI was an industry standard, so lots of expansions would have become available. Just "too late", as always.

Since Commodore have been around since the beginning of personal computing, there should have been a common standard. I mean I think they had enough weight at the time when all this was created and they should be part of the definition of the industry standard. But I'm sure they never try and worked alone. This is where PC compatibility would have make sense. Would have require some height from the management.

Taking about PC compatibility, the real point which really hurt and was very simple to solve, was the capacity to R/W PC floppies out of the box. CrossDos allowed it later. So you would be able to have graphical clones of DOS office softwares and work with your data indifferently on PC or Amiga. I think a lot of person would move to the new visually appealing platform.

It's not dumb characters mode that was needed, nor full PC emulation, it was a bridge with the PC world existing data and it was quiet obvious, even at the time.

Originally Posted by Promilus

Now... C64 doesn't have floppy controller - it uses serial communication with DOS on drive itself (which means floppy controller is OUTSIDE). Serial communication - that one Amiga can handle directly and accurately.

Make me think that simply this part would have been great, the ability to read your C64 floppy data. Would have been a great marketing argument.

[Edit] And at full 1541 speed because the slowness was due a misconception on the 64 side if I remember well. The C64 full emulation card would have come eventually later.

[Thomas Richter]

Originally Posted by Promilus

Well show me platforms with same expansion features, which you can assemble yourself and buy all necessary components in computer store around the corner.

Forget it - nobody cares about hobbyists. It's a small market, and that's rather irrelevant for vendors. That's a niche. You sell masses through pre-configured systems. Boring, stupid office PCs, "good enough" for most people, sold in quantity.

Originally Posted by Promilus

And yet it seems nobody here is able to answer the question why the hell it would've been necessary to put CIAs and SID and ROMs up there...

It wouldn't, because it was irrelevant. C64 compatibility would be just outright stupid. Those that still wanted a C64 would have bought one, and those that did not would not pay extra for such nonsense. It is a completely different architecture - sure, one would have developed a "bridge board" with C64 chips on it, as a second computer within the computer, but apparently, nobody did. Must have had reasons.

[Promilus]

You sell masses through pre-configured systems

Yes, and those units are still based on the very same standards and components.

Those that still wanted a C64 would have bought one, and those that did not would not pay extra for such nonsense.

Well someone did create PC XT emulator for A500 in 1990 - nearly obsolete system by that time. Yeah...

It is a completely different architecture

exactly, like x86 with ISA ... yet Commodore heavily invested in that idea and that's why A2000 and A3000 has ISA slots in-line with Zorro to provide ISA functionality if PC h/w emulator card is inserted. Commodore made a silly thing like Commodore64GS. Yes in 1990 as well. I believe C64 mode in A1000 makes a lot more sense.

Must have had reasons

Like... making C128 and selling it alongside C64 up to 90s? That was fine right? And continuing the sales of A500 when A600 was presented suddenly wasn't.
You're missing the point.

[Thomas Richter]

Originally Posted by Promilus

Well someone did create PC XT emulator for A500 in 1990 - nearly obsolete system by that time. Yeah...

Because the PC *was* relevant for business, the C64 not. See the difference? Professional - hobbyists?

Originally Posted by Promilus

Like... making C128 and selling it alongside C64 up to 90s? That was fine right?

No, nonsense, and another nail in the coffin. The C64 was, at that time, already outdated, and the C128 not a major improvement. The same type of nonsense as the A600.

Originally Posted by Promilus

You're missing the point.

No, you are.

[Promilus]

Originally Posted by Thomas Richter

Because the PC *was* relevant for business, the C64 not. See the difference? Professional - hobbyists?

Yeah... I do see a professional on A500. Are you actually believing that sh*t?

No, nonsense, and another nail in the coffin. The C64 was, at that time, already outdated, and the C128 not a major improvement. The same type of nonsense as the A600.

And still... sold well over next 5 years.

No, you are.

Right. There's only one truth - the one relevant to you. Like MMU.

[Gorf]

Originally Posted by Thomas Richter

Because the PC *was* relevant for business, the C64 not. See the difference? Professional - hobbyists?.

The Sidecar (and the Transformer) did probably more damage than help:
The resources in developing these would have been better spent in polishing Textcraft and creating "Cellcraft" … essentially a good native office suite.
Relying on more or less a whole PC to show the Amiga can do business apps, is only emphasizing on a perceived weakness :-/

An affordable (and optional!) C64-card would have been different, as the advantages of the Amiga over an 8-bit machine were so stark and obvious - it would have clearly been only a tool to ease customer’s transfer to a new platform and showing everyone Commodore cares about compatibility between its product lines and cares about its customers … it would have been good marketing for once

[Gorf]

Originally Posted by chb

The VIC II does not have RGB output, it even does not have YUV output, but luminance and a color signal (3.58 MHz carrier QA-modulated with the I and Q color components in the NTSC case). So either a separate video output or de-modulation to RGB and mixing/switching with the Amiga RGB signal. Or some fancy digitizer (in 1985?), or a redesign of the VIC.

You make it too complicated.
A small redesign (one could even argue a simplification) of the VIC-II would do:
Stay in the digital realm!
At some point in its internal pipeline the VIC-II has a 4-bit value for every pixel.
All you need from there are four dual-ported line buffers.
From the Amiga-side these buffers look like a small region in ChipRAM
(Not entirely unlike what the A314 can do today...)

Then you just need to synchronize the Amiga to the VIC-II-timing and point the bitplane-pointers to these buffers. Et voilà!
Now you have your C64 output as an 16 colour Amiga LowRes screen .. next to other screens and you can even change the palette!