So you were put in charge of making the 060 based amiga

[tone007]

I may be able to offer a different perspective here as someone who'd never seen an Amiga in person until 2007.  Just call me part of the unwashed masses.  I recall 1992-1993 vividly, it was my first year of high school and I started shopping for a machine to replace my Apple IIe, which served me well during the last years of grammar school for printing reports, after my C64s died.  Back in '93, Amiga wasn't on the radar of myself or anyone else that I knew in good old NJ, USA, it was a faint memory from the 80s having seen screen shots on C64 game boxes.  In '93, I saw Windows 3.1 and Wolfenstein, and those were must haves! I didn't get my Windows 3.1 machine then, but I did get an 8088 as it seemed IBM compatibles were the way to go, and I played with DOS and burned up the phone lines with 300 and 2400 baud modems until I got a 486 at some point in 1994.

So basically, I would've had to have offered the '060 based Amiga with PC compatibility as standard, and pushed that point in advertising.

[runequester]

lots of cool and interesting ideas from you guys Keep it up

[XDelusion]

Something better than the AGA chip set.

[nicholas]

1. 060/66Mhz
2. 3DO Chipset on a PCI Card
3. PROFIT!!

[ElPolloDiabl]

You don't need an 060 just an 030 Amiga with lots of fast ram for the low end at least.
For the high end put in a graphics card a 2MB low end a 4MB high end and a super high end card with Nvidia TNT2 level performance.
Also put in a software based Macintosh and PC emulator.

[psxphill]

It must run an 060 processor.

For low end you'd probably have to go for something lower than an 060. Maybe either an 030EC or 040EC.
 
Then:
 
The rumoured aga+ with native chunky pixels.
Add texture mapping to blitter.
Add more sound channels to paula, with proper stereo panning.
Up to 8mb of chip ram.
 
It would be too late by 1996, you'd need to have gotten that out to developers by 1994.

[Hattig]

People seem to be forgetting the price of RAM and computers back in 1996!

8MB of RAM wasn't cheap, it would not have been an option for a consumer system trying to hit £299. Maybe for a higher-spec variant at £499 however...

The same goes for the 68060 versus the 68030. I think a 40MHz 68030 would have been the upper end of what a £299 computer could have borne, price-wise. I don't even know if a hard drive could have been included at that price either. Again, a higher-spec variant could have included it (leading to the concept that the low end system's CPU would have been on a removable card from the beginning).

A1600 @ £299: 40MHz 68030, 4MB (2 + 2)
A1600 @ £499: 40MHz 68030, 8MB, 120MB (I forget the prices of 2.5" hard drives in 1996!)
A1600 @ £699: 50MHz 68060, 8MB, 250MB

The 68040 would have been too hot for the form factor to be cooled quietly. However maybe a 25MHz variant could have been used.

I believe that Commodore missed an opportunity with not releasing "higher spec" consumer models - something for users to aspire to. Why couldn't the A500 also have a better configuration (£100 more) with a 14MHz 68000 and 1MB at the time it came with 512KB?

However you have to consider that this would have been the last hurrah for the 68k line of Amigas, Motorola weren't in the mood for improving the 68060, nor price reducing it. Commodore would have had to be in the process of transitioning to a new architecture by 1996, knowing these machines would be a final 68k release. PA-RISC might have fallen by the wayside. PowerPC would have been a serious consideration. Most likely the C= management would have thought "let's wait for Itanium", and eventually died from the delays and sub-par performance.

[persia]

Make the move to X86....

[Amiga_Nut]

They still would have been screwed come 1995. They had nothing in AAA that would have come close to the Rendition V1000 or hell, even the Matrox Mystique in terms of graphics performance, and sound wise there were some superb consumer grade sound cards coming at this time that were more than a match for the Mary sound chip.

--edit--

Hombre might potentially have held its own if it'd been released in 96 within the games console market, but AAA - Acutiator? It was a dead end.

But for Quake style or solid textured 3D games (all the rage in 95/96) you needed a super fast CPU anyway so it didn't matter really, in 96 you would need an 060 machine to compete with 90/75mhz bargain bucket Pentiums. Hombre would never have been cost effective enough for a console rivalling computer anywhere near A1200 costs and Saturn/PSX was sold at a massive loss for years. Had Commodore committed a large enough order to Motorola for 060s it would have reduced the cost massively, R&D costs for Hombre would have dragged on for 1-2 years after release and even then maybe not viable using a cheap 030/040 CPU.

At least with Ranger we would have had upgrades superior to AGA earlier, ideally instead of the lame ass ECS 'upgrades' and if they had invested in it early enough with a console style 64/32mb VRAM archictecture as a go between Chip and Fast ram it may have been enough to bring in the sales because A500/600 lost a lot of ground to SEGA and Nintendo in 1990/91 due to no decent parallax(2x8 colour playfields was pathetic), only 4 sound channels, no HD floppy drive to reduce disk swapping, only 32 colours as EHB was slow as hell on OCS/ECS, weaker sprites than a C64 in some respects etc etc.

Ranger was done and gathering dust since 88, they probably lost the plans in an office move knowing Commodore!

I think the problem was Commodore, even the C65 prototype managed 320x200x256 colours with a blitter on board in mid 1990...so why did it take them another 2.5 years to get AGA/256 colour mode into an Amiga?

Actually I think the problem in 1992 was the A4000/030. It was just too expensive for a middle range machine, they needed an 020 28mhz 2mb chip & 2mb fast ram machine in 1992 using the A1200 motherboard as is and just shove it in an Atari Mega ST/A1000 style slim case with external keyboard for £499 or £599 with basic slow CD-ROM etc. A 28mhz 020 is pretty much same speed as a 28mhz 030 really and all you pay extra for is case plastics, external keyboard from A4000 and some RAM. I'm sure 499 or 549 was very do-able. But your choices were 1000+ bucks for 4000/030 or crippled A1200 with no edge connector/zorro and cheap $1 keyboard for 399. Hell you could get 33mhz 020s too and probably clock them up to 40mhz with a fan. Bad marketing, they hoped the accidental success of the A500/2000 would repeat itself AFTER PCs had caught up with superior chunky VGA mode and multiple 16bit DACs on board as well as far faster CPUs as standard and PCI bus architecture rivalling the A4000 on every Pentium PC let alone A1200.

[the_leander]

But for Quake style or solid textured 3D games (all the rage in 95/96) you needed a super fast CPU anyway so it didn't matter really, in 96 you would need an 060 machine to compete with 90/75mhz bargain bucket Pentiums.

Right up until the release of the RV1000 hardware 3d version of quake, or a little later the OpenGL release. At which point not only are you being outgunned in terms of raw cpu performance, but now you were being utterly pwnt by the 3d accelerators finally being given something to really chew on.

Hombre would never have been cost effective enough for a console rivalling computer anywhere near A1200 costs and Saturn/PSX was sold at a massive loss for years.

With the exception of the Wii, every single games console since the PS1 (and quite possibly earlier) has been sold at a loss, with the game prices being hiked up to subsidise it. There is no reason why such a tactic could not have been employed for Hombre. Certainly it was significantly less complex on paper than the abortion that was the Saturn.

Actually I think the problem in 1992 was the A4000/030. It was just too expensive for a middle range machine,

The A4000 was too expensive period. It only ever got worse as time went on, from the initial 030 model right up to the AT A4000 with it's 060. Ranger wouldn't have helped all that much. Chipsets of the type offered by the Amiga, even by 1992 were looking like a decidedly bad thing to do. Everything else was going modular. By 1994 PCI was out in force, Pentiums were the new kid on the block. A year later Macs had gone the same way.

The gig was up.

[kolla]

Maybe either an 030EC or 040EC.

I would avoid the EC models and offer fully functional CPUs that can handle the ZorroIII bus that you also want to have available.

[orb85750]

If it was to survive it had to go modular. It had to dump its games console roots.

And how long to stick with our beloved 68K series CPU?

[B00tDisk]

Commodity parts, commodity parts, commodity parts.  AGA "on board", 16 bit audio (even if it meant going with Reveal, Creative, Turtle Beach or Opti).  IDE.  One bay to the left, emtpy (for an additional HD or IDE CD-ROM).  Comes with 2mb Chip RAM, 4mb system.  HD not optional, 80mb is your standard.  This model will street at around $1099 or $1099.

The A1800 would be the same basic model, in a larger Mac Centris style case, a PCI riser with two slots, and an additional HD slot under the CD-ROM bay.  $2099 is my desired price.

Some folks may be wondering why I'm not suggesting a yottabyte of Chipram and AAAAAAAAAAAAAAAAAAAAAAAAA+ chipset capable of displaying the colour out of space or something: My aim is to make these the last "custom chipset" Amiga systems ever.  Putting that old hardware out to pasture.

With that said, if I have the time, I'm also having the OS re-written in a higher level language to make porting to PowerPC as seamless as possible.  The goal is that OS5 (the OS for my two above "dream systems" will be OS4 - more on that in a few) will be entirely portable, so the eventual shift to x86 will also be as seamless as possible.  OS5 will include a sandbox layer to run as many old apps as can be (system breaking shit like those boot off a floppy trackloading pal only eurogames...nnnnnnnotsomuch, no.  Not so much), but will be a new, unique OS unto itself, featuring MP (yes), VM (YES) and other modern resource management tools.

Anyway, the OS4 for the two systems above includes basic network extensions, maybe some rudimentary multi-user capabilities (let's say we store three profiles in some extra space in Kickstart, along with passwords).  I'm also on the horn to the Netscape guys for a quick and dirty port (since, again, I can do what I want).  Maybe throw a few dollars at iD for a DOOM or Wolf3d port, too.

These might not be the greatest seeming options, but this is a transitional system.  My goal is by 2000 to be entirely chipset independent, OS 1.3/2.04/3.1/4.0 are pleasant memories - and that's all.  OS5 and descendants are aggressively upgraded, and pimped out.  I'm also pursuing A/UX as a server OS and beefed up or towerized A1800s to run same.

Devcons, devcons, devcons.  I'm putting free machines in the hands of universities, I've got rock-n-roll tour style buses on MIT, Cambridge and CalTech campuses touting this thing left right and center.  I'm putting out slick flyers with every verbal misstep by Apple and Microsoft in big bold letters - "Bill Gates didn't even mention the internet in The Road Ahead...I guess you took the wrong road, Bill..." "Steve Jobs said '1984 won't be like 1984' - and released a computer that's next to impossible to upgrade as you like.  I guess Big Brother thinks it's OK to bully everybody." et cetera ad infinitum.

My promo videos are showing Wing Commander II and Strike Commander (remember, mid 90's = the golden age of flight sims on the desktop) running...then the screen getting dragged down and we see ol' classic Tut in DeluxePaint VII looking at us.  All the good guys who released great hardware for the old A= boxen from the get-go are being pumped up with free hardware, invites to devcons, and everything else I can think of to get them excited about the new hardware.

We're burying SGI, Apple is barely muddling through by trying to release clones, Microsoft is caught with its collective shorts around its ankles (all the while still releasing Office on our platform!) trying to get their own working browser (which as soon as we see NS frittering their marketshare away, we politely request a port of).  We've ditched our A/UX OS for "Workstations" and servers and have gotten a port of NextStep, and are hungrily eyeing NeXT as a corporate acquisition...

Around Y2k, we're strongly interested in the coming fad of "mp3 players" for the Christmas season.  Maybe this Jobs chap we hired when we bought his company can add some cachet to that project

[Marcb]

With the benefit of hindsight...
 
Get on the Internet bandwagon, built in tcp , Web browser, e-mail clients, newsgroup readers.
Fund multiplayer game development and offer free server hosting to start up game developers.
Infiltrate popular culture with school/sport sponsorships, product placement in popular TV shows.
Drop propietary h/w or open it up to clone makers, make the OS my core business.

[dougal]

Back in 1995/1996 Commodore should have for their low end machines used 3.5" IDE hard drives rather than 2.5". Bigger, Hotter? Yes. But much cheaper. Around that time a 3.5" 1.0GB hard drive was very affordable.

All Amiga's should have had a Buffered IDE as standard.

Even back in 1992 when the A600 and A1200 were released, those 20MB and 40MB hard drives were a joke and badly outdated. By then the standard in a PC was already 120MB.

[Hattig]

Back in 1995/1996 Commodore should have for their low end machines used 3.5" IDE hard drives rather than 2.5". Bigger, Hotter? Yes. But much cheaper. Around that time a 3.5" 1.0GB hard drive was very affordable.
All Amiga's should have had a Buffered IDE as standard.

I think that going for a low-end desktop model with separate keyboard would have been a good idea as well - like the never-released Falcon040 computer from Atari (or an updated A1000, or a slimline case like a PS2). This would have had space for the cheaper 3.5" hard drive (although you could fit certain slimmer models inside the A1200 at a push, I certainly did!) and an internal power supply, but otherwise the innards would be a $600 A1200 rather than a $3700 A4000. It could have sold for $999, and maybe the budget would have squeezed to 4MB RAM and a 28MHz 68030.