Comment by wslh | Hacker News Reader

Comment by wslh | Hacker News Reader

Parent Full thread

My memory is very different about the timing. Not saying that your claim of sustaining innovation is not very complex.

I remember that Microsoft continued with DOS for long and Windows <= 3.x (and a bit later too) was [quasi] just an UI on top of DOS.

My main explanation in 2024 is that Microsoft is Microsoft, no matter what they sell. Microsoft is a business machine like nobody have seen before, they fail at zillions of projects and acquisitions but they continue to exist and growth! Think that Apple found their way very late on so business wise Commodore was doomed, it is not about the hardware. I remember in the late 80s joking to my friends spending money on a Sound Blaster and a CGA... VGA.... SuperVGA. I developed in assembler for the 68x line and just after I learnt some x86 I couldn't believe you need to know about offsets for pointers. But Microsoft is Microsoft and Intel was Intel. The 386 line changed everything. It is documented in the Andy Groove book.

This is not personal with your comment but I think several times in HN the community doesn't understand the complexity about going to market and touching the really complex business reality. This goes far from using fancy programming languages, frameworks, etc. Even receiving huge investments.

>The 386 line changed everything.

The i386 was first offered for sale approximately 12 months after the Amiga launched.

Indeed, based on Wikipedia ~4 months of difference but we are comparing a microprocessor with a full computer though. The Amiga included capabilities (sound, graphics, sprites) that took more time to incorporate in average PCs.

A last observation, I remember sprite support was an important capability for early microcomputers but the PC completely ignored that.

In my view it is entirely about the hardware. By the time the Amiga 1200 came out, PC makers could sell you a 33 MHz 486 with <10 cycle multiply, and VESA graphics card with convenient 1 byte/pixel graphics memory arrangement. Against this, muh blitter and HAM and dual playfield mode means nothing. Your dad's bitplane games might be smoother on the Amiga, meanwhile Wing Commander looks like nothing you have ever seen before and let's not even get on to Wolfenstein 3D. (As a kindness to Amiga fans, I am not even going to mention Doom or Syndicate.) The PC games ended up better, and RIP Amiga.

I am not saying that there was not a breaking point but I say that Amiga had enough time and its demise was not inevitable. This was much before Doom. As it is say in the thread they could see themselves as a game console and market that. Nintendo Switch is worse than PS5 but they remain in the market and they started selling game cards. May be Commodore needed a Mario and a Donkey Kong. I think the three coprocessors in the Amiga were a breakthrough before intel included the FPU. This is why I return to the business execution part. And also Commodore was confronting Microsoft and Intel directly. I also felt that Commodore was a good replacement for the Mac, it has WYSIWYG word processor, 3D software, etc.

As someone who lived through that era, I agree, the 486 murdered almost all of the competition (and nearly even killed Macintosh). Practically everyone I knew who played console games bought a 486 to play games on instead (and also that newfangled Internet thing didn't hurt, which was much nicer on a 486)

Further, Commodore made a big bet on the CD32 but 3DO was already out there, Sega Saturn was coming out and PlayStation was looming (and would have finished Commodore off had it lived another year)

Like Atari Corporation, Commodore failed to innovate enough and the result was not unexpected.

CD32 was not so much of a big bet as a hail mary while hoping to buy enough time for their other chipsets to be completed. They clearly did see a big future there - up to and including the point where their last chipset (Hombre) was planned to be usable either as a standalone chipset for consoles or set-top boxes or as a chipset for a larger computer (or a high end graphics card).

But the CD32 is little more than a CDTV-like variant of the A1200 with a quick and dirty attempt at mitigating the by then disastrous miscalculation of not having any chunky modes in AGA by adding chunk-to-planar conversion in hardware into Akiko (CD32 specific chip that otherwise served to cost-reduce a bunch of glue logic and controllers for the CD-ROM and game ports), but it was very basic (write 32 8-bit chunky pixels to Akiko registers; read 8 32-bit words for 8-bitplanes worth of 32 pixels back out).

> As a kindness to Amiga fans, I am not even going to mention Doom or Syndicate.

I played the hell out of the Syndicate demo on my Amiga, maybe you’re thinking of something else? It was a ton more playable on there than any of the Amiga attempts at Doom clones of the time, thats for sure.

There was definitely a point where PCs left the Amiga in the dust and it was around the time VGA and better started happening though. The A1000 was an amazing beast when it came out but Commodore basically sat on their ass when it came to improving the thing.

> There was definitely a point where PCs left the Amiga in the dust

Yes but by that time you had other home computing platforms like the Acorn Archimedes, arguably the next evolutionary step in the home computing niche. x86 hardware would've been very expensive at the time and mostly used for business purposes, not in the home.

The Acorn Archimedes was never serious competition for Commodore in that space.

I remember - as an Amiga user - cheekily pitching software upgrades to Acorn to help address the massive shortcomings we saw in RiscOS compared to AmigaOS. Even ca. 1996, a couple of years after Commodore's device, we were still smug about how much better we saw the Amiga as being.

We were willing to accept the hardware wasn't bad, but the Amiga userbase never really considered the Archimedes in the same ballpark as an overall system.

I got an Acorn A3010 to replace my Amiga 2000. It was a massive improvement! Not for games as there weren't all that many but the more serious software was amazingly good.

Impression Publisher ran rings around every DTP package on the Amiga. It was an all around better machine, for those kind of tasks. I still miss the Draw program that was included in the ROM (!) of RiscOS.

But alas, the PC made it all history. And while I'm happy with my current hardware, it sure isn't as exiting as the older era...

I'm not surprised - the A3010 was 5 years newer. The more direct competitor for the A3010 would have been the Amiga 1200, but their specs meant they certainly targeted very different markets - "nobody" bought any of the "small" Amigas for DTP.

No, definitely Syndicate! On an Amiga 500, it felt much too slow compared to the PC version. Compared to the 68000, a 1970s design, the higher clock speeds and much better architecture of the 386 and better really helped. (No, you don't have as many registers... it is still a lot more effective than the 68000!)

It also looked a lot nicer as it ran in 640xsomething mode (640x400 or 640x480 I assume).

Perhaps it was a lot better on an A1200. The 68020 is also a much better CPU than the 68000.

You should read up about the corporate shenanigans at Commodore. The company was effectively self-kneecapped by uninterested suits; they would likely not have survived regardless of any development in the field.

Amiga had good 3D games before the PC; breakthroughs like DOOM happened when the company was already on its way down, after a number of wrong strategic turns - including an unwillingness to foster a liberal ecosystem of third-party add-on manufacturers, which was what eventually dragged the IBM-PC into modern multimedia.

By the time of the Amiga 1200, Commodore was effectively already dead. It was a walking corpse.

AGA, the chipset in the A1200, was a stopgap produced because the originally intended graphic chipset, started in 1988, had been beleaguered by delay, in part because of severe underfunding of R&D, and they needed what was effectively a quick and dirty intermediate hack fast.

The "AAA" chipset that was intended to be the next generation, and which had e.g. chunky graphics modes and massively upgraded memory bandwidth and blitter speeds etc., and would've made all the difference for things like Doom, never made it out the door. The project was entirely cancelled in '93, as they concluded PC's would catch up to AAA by the time it'd be finished.

Instead, after AGA, Commodore tried one last ditch attempt and got frustratingly close to getting a massively upgraded design w/3D acceleration, up to 32 bit chunky modes and built to scale from being a standalone system (with an on-chipset PA-RISC CPU) to be licensed for consoles and set top boxes, to a chipset for an upgraded Amiga, with an option to also sell it as a high end graphics card.

[Of course, whether any of that would have saved Commodore was another matter - without a severe overhaul of corporate culture, it was just a question of time; Commodore had basically never, through it's entire existence, even as it rapidly grew in the 8-bit days, been more than one unresolved crisis away from collapse]

Reply on news.ycombinator.com