The last two paragraphs make it sound like the C64 was the last successful product that Commodore made before imploding, and claim that Commodore never recovered after Tramiel left in early 1984.
But, as anyone who was around then knows, this is not the case, because the most successful Amiga model, the Amiga 500, came out three years later, in 1987.
Saying that Commodore was finished in 1984 seems like a gross oversimplification at best.
The C64 was a money spinner from Commodore all through the 80s and early 90s. It's their Apple 2 which kept Apple afloat with the grandiose Mac boat anchor which didn't actually make any money for Apple until Steve came back. The Amiga range was never really profitable for Commodore the way it should have and Jack would have released the A600 in 1985 which is exactly what was needed then.
Tramiel and his sons and a big portion of the team around him at Commodore had their chance to make rock bottom price 16-bit machines after they left Commodore. That's what the Atari ST was.
I loved mine, but it was no C64 in terms of market success. While it initially outsold the Amiga (because it came out earlier and cheaper) it plateaued and then lagged behind once the A500 came out. And both platforms hardly sold anything at all, in the grand scheme of things.
The mid and late 80s home computer market was a very different place from the early 80s. His product development and sales style didn't do him any favours from 1985 on.
built-in MIDI ports,
low price,
high quality paper white monochrome monitor,
mac-like GUI,
tight low latency timing (single tasking OS until the early 90s -- straight from the interrupt to the subroutine do not pass go).
Cubase and Logic both had their origin there and are still around.
Myself, my mom and dad wouldn't spring for synthesizers for me, despite my mom being a music teacher. So I collect them (and Ataris) now but they just collect dust :-)
I don't recall the exact crossover point, but the Mac starting making serious money circa 1987 and the Apple II became irrelevant in terms of bottom line not long after. Steve came back in 1997.
The article completely omits any mention of the Atari 8-bit computers which used 6502. It skips from Atari VCS/2600 to C64.
Atari 400/800 GTIA & ANTIC were a massive improvement over the TIA in the VCS and came to market in 1979 more than 2 years before C64.
Atari 8-bit had better colour than the C64 (palette of 128). C64 had much better sprites, but Al Charpentier had the benefit of seeing what had been already been done by Atari.
Sure, let's shoehorn in lots of random computers from overseas. BBCs only sold to schools and education but not to most parents as they were too expensive. ZX Spectrums were Z80 based which isn't a 6502 processor. Apple 2 is mentioned in the article.
I think the point is that they sold fairly well compared to other successful machines of that era. They probably also outsold a number of popular game consoles as well such as Atari's 7800.
Not to mention that the Ataris were way earlier, and were at one point regarded as the MOST popular computer worldwide. That's why the learning platform PLATO was released on the Atari first.
I don't think the historical importance has anything to do with sales numbers. The article mentions Commodore PET, of which only 200,000+ units were manufactured.
The Atari 400 was a much bigger technological achievement than the Commodore 8-bitters that were released later though. It deserves its place in history.
(it also was the predecessor of the greatest computer of all time: the Commodore Amiga ;P)
Yeah, it's a baffling omission that undermines the whole thing. The Atari 8-bits were pretty incredible, combining the 6502 with not only a way better graphics co-processor than the VCS, but also dedicated sound and I/O chips.
The Atari 800 had S-video output, in 1978! Come on, that's pretty forward-thinking. I still have my 800 and Commodore (JVC) 1702 monitor, a state-of-the-art combo in its day.
The article makes it sounds like it was difficult to update the AppleII screen so that there was no flicker. It was not difficult. We had plenty of time.
I seem to recall we could flip between two different high res screens and so we drew on the hidden one and then flipped to the other one when the vertical blank interrupt occurred. Also, we rarely had to update the entire screen all at once.
Both the Apple and C64 had awkward graphics, but graphics speed was not a concern. The issue was more often disk storage and access speed, and of course memory limitations.
The thing I am most proud of as a programmer is I wrote an emulator of AppleII graphics for the C64 so we could reuse the art from the Apple version of the game. It was just barely fast enough.
It was the Fisher-Price series. Peter Rabbit Reading and Man in the Moon Math. I also did Alphabet Zoo, Adventure Creator, Hey Diddle Diddle, and maybe others I am forgetting.
Mmmm, I don't think so. The Apple had only blitting, whereas the Atari, Commodore, and TI had hardware sprites.
And I have no idea what you mean by "C64 had awkward graphics." It integrated ideas already present in the Atari computers, with some advantages and some disadvantages. Both those platforms utterly crushed the primitive graphics and "sound" available on the Apples, with their hardware sprites, fine scrolling, redefinable character sets, and multichannel sound. The Ataris also used intelligent, daisy-chainable peripherals.
Ugh. When I was in junior high and going to the school library to use the Apples, of course I wanted to write games. I kept asking the teacher how to make moving graphics. Needless to say, nobody knew.
At some point I got clued into shape tables, plotted one out, and then called upon it in a program. It was no faster than a sequence "drawto" statements. At that point, I had my answer: You don't; at least not on an Apple.
The scrolling model used on the c64 means you have to copy the entire screen which is difficult to do quickly enough to avoid glitches. In contrast for instance the NES outputs the content of a sliding window on a large buffer so you only have to move the window and draw the newly exposed screen area. That way you can create more dynamic games with the same processor.
I seriously doubt that. This certainly wasn't true on the Atari. The Atari had fine scrolling and no need even to redraw newly exposed areas. The screen was, as you describe, merely a window upon as much memory as you had available.
Because the Ataris and Commodores had redefinable character sets, you could create graphic elements as characters and build huge maps and scroll around them smoothly with ease.
It’s a 1mhz processor and for scrolling to be smooth you need to move 2k of memory within one frame. That means if you get the timing right you have at most 1 million/60 clock ticks, which is about 16666, to do the scrolling and everything else you need to do.
You really can’t do that ‘with ease’ which is why c64 games typically divide up the screen so the scrolling area is smaller.
You don't move the memory at all. You move the window. If that's not clear at this point in the discussion (let alone 30 years later) I don't know what to tell you.
It by “blitting” you mean the Apple had a single 6502 chip which could move bytes around, then yes. But it had no dedicated blitting chip like the Amiga had (which I never got to work).
The Atari graphics were awkward, too, or so I was told by my boss at the time who did the Atari versions of our products. Good though. What I mean by awkward is just that I was surprised, as someone with no training in computer engineering, how many little quirks you had to work around. Apple graphics were primitive compared to the others, using a seven bit architecture for pixels and the high bit used for color information. At least the C64 had all eight bits for pixels, which made rendering a little faster. C64 sprites were interesting, but like the OP said we had to do tricks with interrupts to reuse them as the screen rastered.
I guess another reason I said awkward is that we had to do every trick we could to squeeze the art into memory. Alphabet Zoo was shipped as a 16k cartridge. I did an entire game with art in 16 kilobytes! I used nine bytes to encode my name, then later I needed those nine back again… so no easter egg for me.
Blitting refers to rewriting bytes, yes. It does not imply dedicated hardware to do so, which is why Apple games flickered.
Calling Atari and Commodore graphics "awkward" in the scenario you describe sounds like laziness in learning and taking advantage of new graphics hardware. The sole "awkward" thing on the Atari was that, for some reason, the sprites were as tall as the screen and therefore the only way to achieve vertical movement of a sprite was indeed to move it through memory. But you could do this even in BASIC by setting the sprite data up in a character string and using the fast string functions to move it.
Both the Atari and Commodore made graphics programming decidedly non-awkward by providing automated collision detection between the sprites, and also with the background. Moving sprites around during screen refresh was common on both platforms. On the Atari, the technique (called a display-list interrupt) could also be used to put 256 colors on the screen at once and simulate multi-color sprites.
An innovative use of display-list interrupts gave rise to this cool effect that was called (by LucasArts if I recall) "altered-perspective scrolling." Check it out here in BallBlazer, a soccer-type game written for the Ataris: https://youtu.be/_tkwWD_BWWQ?t=631
No, Apple games did not flicker. We used screen swapping to stop that. It's the first lesson I learned as a video game programmer: flicker not allowed.
I don't know what your deal is with the word "awkward." As someone who wrote a lot of games on Apple, C64, and one each on the PC and Amiga, I've earned the right to use whatever innocuous word I want to describe a vague feeling that I have. Mind you, all that more than 35 years ago. Perhaps I am misremembering how non-awkward it all was. I did not do anything with the Atari. I was merely told by my boss and friend Dale Disharoon how strange he found it.
Yes, we liked the C64 collision detection. The main awkwardness about the C64 was, at the time I was first working with it, that it was all undocumented. I just think it's awkward to call information to get the phone number of a writer of a magazine article in Toronto to find out how to work the vertical blank interrupt, since there was no reference manual at that time. The Amiga manuals were full of errors when they first came out.
I suppose I am a lazy learner. As you can read all about in my book about being a high school dropout, published by Simon and Schuster (get your copy in Korean today!). Also see page 60 of this issue of Compute Gazette (https://archive.org/details/1983-11-computegazette/page/n59/...).
Apple II games didn't flicker, but exhibited tearing because there was no indication of the vertical retrace: no I/O pin or interrupt, nothing.
I don't remember a lot either and am searching for this now. It looks like the Apple IIe added a bit at $C019 for detecting vertical retrace, but the II+ and earlier didn't have it.
It sounds more like you want to say that the Apple II+ sucked. That’s fair. I don’t disagree with that. The sound was generated by a single click which if repeated fast enough became a buzz. I wrote a nice little routine to emulate pink noise, once, so it would sound like a car skidding. That was kind of cool.
I sense that you are too young to have first hand knowledge of these things, but if you do then your memories are just as faded as mine. I remember that we didn’t have flicker. And I was in the industry at the time developing games for the Apple II.
Author seems to tie things off after the C64 era, but the 6502 had a lot more to give. The penultimate video gaming on the 6502 is IMHO the Atari Lynx. 6502 + almost Amiga-esque (David Morse of Amiga fame was involved) custom graphics processor that could do 4096 colours, unlimited sprites, hardware multipliers and dividers, bunch of custom processing stuff to augment the CPU
The combination of the excellent interrupt responsiveness of the 6502 with a display chip of that caliber is pretty cool. Too bad it wasn't a market success really.
The SNES is also pretty damned amazing (yeah, 65816, but close enough)
The 65816 is a fascinating chip, even though it's a dead end in the evolution of home computing. It falls in a category with the Z8000 and maybe even the 80286: backwards compatible with their predecessors, but extending them in odd, idiosyncratic ways. Processors evolved and mostly discarded the vestigial weirdness of the original 8-bits, but its a special kind of fun to revisit assembly language programming in that style with some of the oppressive constraints (address space, addressing modes, minimal registers) lifted.
The original Game Boy had some very interesting capabilities as well. In some aspects it outclasses the NES, in some others it lags behind.
(I specifically mean that the NES only had sprites and a single background layer. The Game Boy had an additional 'window' layer for drawing eg status bars, but also capable of doing funky things in eg boss battles. It also got some capabilities built-in that the NES needed extra chips on the cartridge for.)
Looking at what the Game Boy could do, you get the sense that they carefully optimized for the exact specialised capabilities they needed for games at minimal transistor-count cost. Being such a late 8-bit system helped a lot with that.
> But as with all semiconductor products, the price of the 68000 plummetted as volume production began, and it now [in 1984] sells for about $15.
The 6502 evolved into the 65C816 but those price advantages relative to processor like the 68000 had a way of evaporating into thin air. If there’s anything that explained the “fall” of 6502 gaming, it’s the fact that the processor never really made it to the 32 bit era, and the 32-bit processors fell in price to the point that you weren’t saving money by using the 6502 or 65C816 in your designs.
39 comments
[ 3.3 ms ] story [ 132 ms ] threadBut, as anyone who was around then knows, this is not the case, because the most successful Amiga model, the Amiga 500, came out three years later, in 1987.
Saying that Commodore was finished in 1984 seems like a gross oversimplification at best.
I loved mine, but it was no C64 in terms of market success. While it initially outsold the Amiga (because it came out earlier and cheaper) it plateaued and then lagged behind once the A500 came out. And both platforms hardly sold anything at all, in the grand scheme of things.
The mid and late 80s home computer market was a very different place from the early 80s. His product development and sales style didn't do him any favours from 1985 on.
built-in MIDI ports, low price, high quality paper white monochrome monitor, mac-like GUI, tight low latency timing (single tasking OS until the early 90s -- straight from the interrupt to the subroutine do not pass go).
Cubase and Logic both had their origin there and are still around.
Myself, my mom and dad wouldn't spring for synthesizers for me, despite my mom being a music teacher. So I collect them (and Ataris) now but they just collect dust :-)
Atari 400/800 GTIA & ANTIC were a massive improvement over the TIA in the VCS and came to market in 1979 more than 2 years before C64.
Atari 8-bit had better colour than the C64 (palette of 128). C64 had much better sprites, but Al Charpentier had the benefit of seeing what had been already been done by Atari.
Four million units for the Ataris in the U.S. versus 5.7 million for the C-128 worldwide.
Seems pretty significant to me, especially compared with other 8-bit computers that people are always talking about on HN:
- BBC Micro: 1.5 million
- Sinclair ZX: 5 million
- Apple ][: 4 million
(it also was the predecessor of the greatest computer of all time: the Commodore Amiga ;P)
The Atari 800 had S-video output, in 1978! Come on, that's pretty forward-thinking. I still have my 800 and Commodore (JVC) 1702 monitor, a state-of-the-art combo in its day.
I seem to recall we could flip between two different high res screens and so we drew on the hidden one and then flipped to the other one when the vertical blank interrupt occurred. Also, we rarely had to update the entire screen all at once.
Both the Apple and C64 had awkward graphics, but graphics speed was not a concern. The issue was more often disk storage and access speed, and of course memory limitations.
The thing I am most proud of as a programmer is I wrote an emulator of AppleII graphics for the C64 so we could reuse the art from the Apple version of the game. It was just barely fast enough.
And I have no idea what you mean by "C64 had awkward graphics." It integrated ideas already present in the Atari computers, with some advantages and some disadvantages. Both those platforms utterly crushed the primitive graphics and "sound" available on the Apples, with their hardware sprites, fine scrolling, redefinable character sets, and multichannel sound. The Ataris also used intelligent, daisy-chainable peripherals.
https://en.wikipedia.org/wiki/Shape_table
At some point I got clued into shape tables, plotted one out, and then called upon it in a program. It was no faster than a sequence "drawto" statements. At that point, I had my answer: You don't; at least not on an Apple.
Because the Ataris and Commodores had redefinable character sets, you could create graphic elements as characters and build huge maps and scroll around them smoothly with ease.
You really can’t do that ‘with ease’ which is why c64 games typically divide up the screen so the scrolling area is smaller.
The Atari graphics were awkward, too, or so I was told by my boss at the time who did the Atari versions of our products. Good though. What I mean by awkward is just that I was surprised, as someone with no training in computer engineering, how many little quirks you had to work around. Apple graphics were primitive compared to the others, using a seven bit architecture for pixels and the high bit used for color information. At least the C64 had all eight bits for pixels, which made rendering a little faster. C64 sprites were interesting, but like the OP said we had to do tricks with interrupts to reuse them as the screen rastered.
I guess another reason I said awkward is that we had to do every trick we could to squeeze the art into memory. Alphabet Zoo was shipped as a 16k cartridge. I did an entire game with art in 16 kilobytes! I used nine bytes to encode my name, then later I needed those nine back again… so no easter egg for me.
But, moving 16bit words was a huge improvement compared to the C64 for obvious reasons.
The one thing the C64 missed was full screen hardware scrolling. I don’t mean the pixel shifting.
I’m aware there’s a clever hack to get full screen scrolling - even at Sonic the Hedgehog rates.
This game leverages that: https://m.youtube.com/watch?v=ldo2ewLBt3Y
Calling Atari and Commodore graphics "awkward" in the scenario you describe sounds like laziness in learning and taking advantage of new graphics hardware. The sole "awkward" thing on the Atari was that, for some reason, the sprites were as tall as the screen and therefore the only way to achieve vertical movement of a sprite was indeed to move it through memory. But you could do this even in BASIC by setting the sprite data up in a character string and using the fast string functions to move it.
Both the Atari and Commodore made graphics programming decidedly non-awkward by providing automated collision detection between the sprites, and also with the background. Moving sprites around during screen refresh was common on both platforms. On the Atari, the technique (called a display-list interrupt) could also be used to put 256 colors on the screen at once and simulate multi-color sprites.
An innovative use of display-list interrupts gave rise to this cool effect that was called (by LucasArts if I recall) "altered-perspective scrolling." Check it out here in BallBlazer, a soccer-type game written for the Ataris: https://youtu.be/_tkwWD_BWWQ?t=631
I don't know what your deal is with the word "awkward." As someone who wrote a lot of games on Apple, C64, and one each on the PC and Amiga, I've earned the right to use whatever innocuous word I want to describe a vague feeling that I have. Mind you, all that more than 35 years ago. Perhaps I am misremembering how non-awkward it all was. I did not do anything with the Atari. I was merely told by my boss and friend Dale Disharoon how strange he found it.
Yes, we liked the C64 collision detection. The main awkwardness about the C64 was, at the time I was first working with it, that it was all undocumented. I just think it's awkward to call information to get the phone number of a writer of a magazine article in Toronto to find out how to work the vertical blank interrupt, since there was no reference manual at that time. The Amiga manuals were full of errors when they first came out.
I suppose I am a lazy learner. As you can read all about in my book about being a high school dropout, published by Simon and Schuster (get your copy in Korean today!). Also see page 60 of this issue of Compute Gazette (https://archive.org/details/1983-11-computegazette/page/n59/...).
I don't remember a lot either and am searching for this now. It looks like the Apple IIe added a bit at $C019 for detecting vertical retrace, but the II+ and earlier didn't have it.
Also: The sound of a planet blowing up on an Apple resembled someone spilling a bowl of pistachio shells.
I sense that you are too young to have first hand knowledge of these things, but if you do then your memories are just as faded as mine. I remember that we didn’t have flicker. And I was in the industry at the time developing games for the Apple II.
The combination of the excellent interrupt responsiveness of the 6502 with a display chip of that caliber is pretty cool. Too bad it wasn't a market success really.
The SNES is also pretty damned amazing (yeah, 65816, but close enough)
(I specifically mean that the NES only had sprites and a single background layer. The Game Boy had an additional 'window' layer for drawing eg status bars, but also capable of doing funky things in eg boss battles. It also got some capabilities built-in that the NES needed extra chips on the cartridge for.)
Looking at what the Game Boy could do, you get the sense that they carefully optimized for the exact specialised capabilities they needed for games at minimal transistor-count cost. Being such a late 8-bit system helped a lot with that.
https://www.nytimes.com/1984/06/29/business/motorola-s-power...
> But as with all semiconductor products, the price of the 68000 plummetted as volume production began, and it now [in 1984] sells for about $15.
The 6502 evolved into the 65C816 but those price advantages relative to processor like the 68000 had a way of evaporating into thin air. If there’s anything that explained the “fall” of 6502 gaming, it’s the fact that the processor never really made it to the 32 bit era, and the 32-bit processors fell in price to the point that you weren’t saving money by using the 6502 or 65C816 in your designs.