What spec would the A1200 had launched with if it was up to YOU

[rothers]

I would have kept everything the same but added a graphics feature to zoom blits.

I think someone asked for this very feature back in the early 1990s on a newsgroup and got a reply (from Commodore) saying no, use the CPU for that.

If the A1200 could have zoomed blits (like 1980s sega arcade hardware for example) it could have run Doom easily... or any ray-casting game.

That would have saved Commodore IMO.

[abu_the_monkey]

a slim line desktop stackable like the original Loraine concept art

bottom unit as an ultra cost reduced a4000 with 020 and 2+2meg ram, small boot HDD, no zorro
top unit is an add on (sold separately) to increase capabilities of the system, zorro 3, 030/040, more ram, extra HDD, etc. probably should have added a 5 1/4" drive bay.

[Thomas Richter]

Originally Posted by Amiga Lifestyle

I feel like the obvious answer is that the 1200 should have matched roughly what Mac's were shipping with at the time.

For that, the Amiga should have had the customers of the Mac, too, that were able to pay the price for it. However, Amiga users were mostly gamers and not power users, and they couldn't afford a MAC (or they would have bought one). CBM missed the opportunity to find a market for their machine, and that was an early mistake. There was a lack of productivity software to attract such users.

[Dunny]

Originally Posted by abu_the_monkey

@Dunny

just for you

droooool

[abu_the_monkey]

Originally Posted by Thomas Richter

For that, the Amiga should have had the customers of the Mac, too, that were able to pay the price for it. However, Amiga users were mostly gamers and not power users, and they couldn't afford a MAC (or they would have bought one). CBM missed the opportunity to find a market for their machine, and that was an early mistake. There was a lack of productivity software to attract such users.

this is true. there was no direction and a whole bunch of odd decisions from the management. kind of just falling into the games machine bracket because of this.

[desiv]

If it wouldn't have increased the price too much, I agree on a 68EC030 at 25 MHz, socketed so it could be replaced with a full 030 with some jumpers on the MB for some faster Mhz 030s.

I would be OK with just the 2M CHIP RAM, but I would have included a couple of SIMM slots for some RAM upgrades, maybe up to 8M.

And the socket installed for the FPU in the empty socket location now...

That would have made basic upgrades much easier.
And then the accelerator slot could be used for 040s and more RAM for people who could afford/want that.

However, I wouldn't have done that if it made the price too much higher.
I could barely afford the 1200 when I got it back in the day.

One of the things I liked about the PCs I also had at the time was that it was pretty easy to upgrade them over time. I was able to upgrade CPUs and RAM as I could afford them, or thru hand-me-downs (I would help people upgrade their PCs, and they would let me keep their older parts).

But for example, the M1230 CPU accelerator about a year after I got my 1200 was $350... That was for a 68EC030 at 40Mhz and just 1M RAM. (Got that from an Amigaworld ad.)
That was a chunk of change for me back then...
It would be much easier for me to add things in smaller bits. (Just some SIMMs for under a couple of hundred, check. Just a faster CPU for under a couple of hundred six months after that, check. But both in a new board to hold them for one chunk of change...)

[Gorf]

Originally Posted by abu_the_monkey

a slim line desktop stackable like the original Loraine concept art

bottom unit as an ultra cost reduced a4000 with 020 and 2+2meg ram, small boot HDD, no zorro
top unit is an add on (sold separately) to increase capabilities of the system, zorro 3, 030/040, more ram, extra HDD, etc. probably should have added a 5 1/4" drive bay.

I very much like that idea und the design!

[Karlos]

Cough *RISC PC* cough

[Bruce Abbott]

Originally Posted by Samurai_Crow

I was referring to the existing belly slot, not a SIMM card on the morherboard.

I know, I was simply pointing out that it wasn't just a matter of Commodore being 'cheap'.

As for including a trapdoor RAM board with the A1200, they didn't do that with other Amigas so why do it now? They could have produced their own expansion board, but the A500 market showed that 3rd parties would soon be making them cheaper. And 2MB was heaps anyway, at least initially. Few Amiga games were demanding even 1MB before then.

Microbotics had the MBX 1200 out in early 1993, and Phase 5 released the Blizzard 1200 / 4 at about the same time. The BSC AlfaRam 1200 (which I sold in my shop) was also released in 1993. By 1994 you could get an Amitek Hawk with 1MB for £99 in the UK.

[desiv]

Originally Posted by Bruce Abbott

As for including a trapdoor RAM board with the A1200, they didn't do that with other Amigas so why do it now?.

Well, with the A500, 512k on release was plenty, so no real need.
Amiga 500 Plus, again they upped it to 1M, so no real need as most A500 games ran fine with only 1M.

And one could say that doubling it again for the A1200 was the same thing and plenty.
Except that we know that at least "some" FAST RAM REALLY helped the A1200.

And yeah, it helped the 500 and others too, but it didn't seem as meaningful a bump as adding FAST RAM to the 1200 would have been.

That said, I am OK with not having done that, as long as they would have made it cheaper to do but just putting a SIMM slot on the MB.

Yes, there were "relatively" inexpensive RAM boards, like the MBX 1200(z), but those were near $180 (USD) without any RAM in them.
I mean, I suppose they could have included an unpopulated RAM board, but it should have been much cheaper to just pop a SIMM slot on the MB.
Having SIMM slots on MBs at the time was pretty standard. Both PCs and Macs were doing that...

[Thomas Richter]

The problem actually started earlier, and the A1200 was already too late. There were plenty of missed opportunities before that, and they started with ECS. The integration of components into Agnus for cost reduction were great, but what the chipset delivered with the upgrade was not sufficient, or nothing the average user could take advantage of.

ECS had programmable video timing (nice) and the possibility to create a VGA signal (or almost an VGA signal) with the 35ns mode - productivity namely. Unfortunately, it offered little else. Ok, larger blits. Nice, thank you...

If you think about it, there were sufficient DMA slots available to have 8 lo-res planes, so why weren't they used? Yes, more color registers in Denise cost chip estate, so that wasn't done, but there were plenty other possibilities. An xEHB mode, 32 colors in 8 luminances would have offered colorful pictures without HAM restrictions, or an xHAM-Mode with 7 planes and 32 base colors. Dual-playfield with two planes at 16 colors each. I believe that would have been more attractive to the average Amiga user than a productivity mode with its limited palette.

The problem was ultimately DMA bandwidth at the chipset side. You would need to double to bus to 32bit to enlarge the bandwith, but that requires a major redesign of the chips, and there wasn't time for it.

AGA as such was not a bad extension, it was just two years too late. A1200 specs aren't too bad, just too late. Having this machine (or a corresponding desktop) two years earlier would have helped.

[Toni Galvez]

It would have been a great publicity move to have made a bundle with the A1200 and the game Doom. An all-in-one, with a proper configuration for Doom to run, because we all had Doom for the Amiga in 1998, even though Jon Romero said that Doom couldn't be made on the Amiga. If this port of Doom had been made for the Amiga in 1993 and sold in a bundle with the A1200, the history of the Amiga might have been different.

My point of view for the Amiga configuration on 1992:

- full 030 CPU at 50Mhz
- 2 MB chip ram with space to extra 2 MB update.
- 2 MB fast memory.
- 8 Channels 16bit sound with Full Stereo panning.
- 80 MB hard Drive.
- High Density floppy drive.
- AGA video with Akiko chunky to planar conversion.

[TEG]

Originally Posted by Thomas Richter

The problem actually started earlier, and the A1200 was already too late. There were plenty of missed opportunities before that, and they started with ECS. The integration of components into Agnus for cost reduction were great, but what the chipset delivered with the upgrade was not sufficient, or nothing the average user could take advantage of.

ECS had programmable video timing (nice) and the possibility to create a VGA signal (or almost an VGA signal) with the 35ns mode - productivity namely. Unfortunately, it offered little else. Ok, larger blits. Nice, thank you...

If you think about it, there were sufficient DMA slots available to have 8 lo-res planes, so why weren't they used? Yes, more color registers in Denise cost chip estate, so that wasn't done, but there were plenty other possibilities. An xEHB mode, 32 colors in 8 luminances would have offered colorful pictures without HAM restrictions, or an xHAM-Mode with 7 planes and 32 base colors. Dual-playfield with two planes at 16 colors each. I believe that would have been more attractive to the average Amiga user than a productivity mode with its limited palette.

The problem was ultimately DMA bandwidth at the chipset side. You would need to double to bus to 32bit to enlarge the bandwith, but that requires a major redesign of the chips, and there wasn't time for it.

AGA as such was not a bad extension, it was just two years too late. A1200 specs aren't too bad, just too late. Having this machine (or a corresponding desktop) two years earlier would have helped.

And so I wonder if all that stagnation was not due to problems with CSG chips production after all. We should examine the timeline evolution of their process. Perhaps they were at the threshold of the transistors numbers they did know to put in a wafer at a reasonable cost.

Perhaps larger wafers would required new costly machines for the production with a good yield (good yield = not too much bad pieces) and the managements did not wanted to invest, just asking the CGS team to do the best of what they can do with the existing process and so limiting the Amiga R&D ambitions (the "read my lips, no new chips" mandate).

This is an aspect of Commodore we should really explore I think to perhaps better able to release our collective PTSD

[buzzybee]

Very interesting to follow this discussion, with its focus on hardware almost exclusively.

But software drives hardware. That was true in the early days of gaming and computing, and it still is to this day. Amiga has always been about games. It lost support from business-driven gaming devs and companies very earl in the 90s. Neither an optimized pixel bandwidth nor a 25 Mhz CPU would have changed that, if not framed by talented developers who use that stuff.

Commodore should have talked to devs from 1988 onwards about what they expected from a next-gen Amiga; never waste time and disappoint people with the A600, but release an A1200 with easier-to-handle AGA chipset including some chunky pixel mode; a game controller; a HD diskdrive; a small hard drive - all accompanied with five exclusive launch games. That might have wowed the audience.

[grond]

Originally Posted by TEG

And so I wonder if all that stagnation was not due to problems with CSG chips production after all.

It certainly was! Just remember that up into the Macintosh era Apple and quite a few other computer companies used microprocessors produced by MOS (=CSG). Why wasn't it possible to build a healthy microchip business on this that was keeping up with the progress of semiconductor technology?

[Thomas Richter]

Originally Posted by TEG

And so I wonder if all that stagnation was not due to problems with CSG chips production after all.

It was, or rather it was along with the entire organization of the company. CBM was a vertical monopoly, they made everything themselves: The chips, the computers, the operating system, all from within CBM. That made a lot of sense in the home computer area since CBM did not depend on anyone, and had full control over their product. However, with the continuous innovation in the field, it became a burden: Other companies became better at making chips for a better price, yet other at producing software, and yet more at designing hardware components from available chips. CBM had to carry all the legacy around, without actually being competative in either field.


That problem, however, was made a lot earlier, by Tramiel.

[oscar_ates]

020 was not a fit to 1992, it should have been 030 by then. Also HD+2MB fast ram on default machine was needed. There was little software available to accelerated Amiga's because of the large user base was using bare configuration. So it was very important that the base Amiga configuration should have been competitive to a 386 PC at the time. It would also not chase away the US developers from Amiga which happened around 1992 because of the lacking harddisk.

Not all people are short in money, all middle class moved to the PCs because they can pay extra money and get the best performance.

Commodore went cheap and put all the burden to the user. 2.5inch hd direct support saved them few bucks for the case and the power supply and cost fortune to the user. That is why it failed to hold foot on the market. Also no simms slot on the base config saved them dram controller etc. again cost $$ to the end user.

[alexh]

Originally Posted by Gorf

Double of what? Double of $10 or double of $50?
What matters here would be the absolute price difference and not the relative one...

68030@25 = $43 vs 68EC020@14 = $15 so it is actually 3x (using prices in the article you linked to).

Originally Posted by Gorf

"$15 for the 16MHz 68EC020, $19 for the 25MHz version"
link So for only $4 more per unit they could have doubled the clock.

Could they though? I think there must have been a dependence with the PAL 7MHz crystal? Otherwise why did they choose 14MHz? By running the CPU at a multiple of the PAL frequency they save a crystal? Simplify the PCB? Are able to run synchronous / faster?

why exactly would 4MB ChipRAM (one more address line) have doubled the size of Gayle*? That does not make sense to me. Could you elaborate?

You are probably right. It's one more address line if you could get chips that are 2x the density. I don't know if that was practical? Otherwise would you need another output enable too? Internal address decoding? The power to be able to drive another bank of devices on the address & data lines? If it was "cheap" they would have over-engineered the silicon with the idea of expanding to 4MB or 8MB in the future?

[TEG]

Originally Posted by Thomas Richter

That problem, however, was made a lot earlier, by Tramiel.

I would not be so affirmative. We know from CBM hardware engineers that they were happy to benefit from short cycles of prototyping and so be able to build new chips or new version of chips quickly. So for a moment CBM had an edge and we would perhaps not have the C64 and the Amiga as we know it, if not. But, as you said, it then became a burden.

I guess, even at the time, it was obvious that the lithography process was evolving fast or will, the Moore law is from 1965. So at a management level, either you decide to sell the activity to an actor able to make it grow and you outsource or you decide to keep-up. Essentially because you know investors will follow you as they trust you and they see the opportunities you will become a major actor in a major field. Gould was a businessman after all.

And MOS tech (then CSG) provided a big part of cpu used in micro-computers at a moment. So I guess they had credibility and so they were not forced to finish in a dead end, the right decisions being made.

No time to check now but I wonder if we have a graph showing market share for MOS/Intel/Motorola and how it evolved.

P.S.
Another possibility would be to eventually merge the activity with Motorola, I mean inventing a deal to keep the edge. Something on this line.

[Promilus]

Originally Posted by grond

It certainly was! Just remember that up into the Macintosh era Apple and quite a few other computer companies used microprocessors produced by MOS (=CSG). Why wasn't it possible to build a healthy microchip business on this that was keeping up with the progress of semiconductor technology?

Apple IIe and so on used 6502 by WDC and other companies (probably WDC licensed). That is the point - MOS actually hardly did produce under Commodore for foreign entities!