Hacker News Reader

There are markets where I think he'd be broadly correct--the TI-83 (well, TI-84 now, I guess) is the one that comes to mind. Its target market gives it a certain cap of feature development (you don't want it too be powerful a computer, lest teachers feel uncomfortable letting students use it on tests), and this means that updating models is largely about minor hardware stuff (e.g., switching out a bank of AAA batteries for a rechargeable battery).

However, given the context of the interview in general ("you're losing market share, sales are slowing, you're losing customers, etc., what are you going to do about it?"), arguing basically "nothing's wrong" bespeaks a certain arrogance that would, even without the benefit of hindsight, give me tremendous pause.

Many people here seems to miss what 95% of users did with the C64 (and other 8-bit machines): Videogames.

The trick was to sell at parents the idea that they were buying a powerful but still cheap machine for their kids so they could learn how to use computers[0].

But once in their bedrooms, 95% of kids spent all of their time just playing games and ignoring what "RAM, bytes, bits" or even Basic was.

Once more sophisticated consoles (Nintendo, Sega) and 16/32bit PC games hit the market, the C64 lost all of its appeal.

[0]: https://substackcdn.com/image/fetch/f_auto,q_auto:good,fl_pr...

That was the starting point, sure, but by the late 80's and into the 90's (production didn't cease until Commodore's bankruptcy in '94) it was still in the game because it was a cheap toy.

I don't have time to go looking for prices in '94 when production ceased, but even as prices for other machines were dropping rapidly, C64 prices remained lower. An issue of RUN from '92 has ads for the C64 at $140-$150. At that point the 1541 disk drive was the most expensive part, at <$165, but you could get the 3.5" 1581 for $109.95

For comparison, an Amiga 600 cost $399 from the same retailer, the same magazine has PC's starting at $500+. Far more capable at that point, sure, but you'd end up paying more than that if you wanted decent graphics and sound, and you needed a monitor whereas the C64 could connect to your TV, and suddenly the C64 was sufficiently cheaper at that point to be seen as a toy.

By late 80s the only "enthusiast" C64 users were mostly (and sadly) those who could not afford a better computer. And from what I can remember, even as a kid, that was not so uncommon.

After all, "a computer on every desk and in every home" was still just a company vision statement.

As an "enthusiast" C64 user in the late 80's, I don't agree. That may have been the case some places, especially in the US where Commodore's sales were dire (not just for the C64). Clearly you're right that some people picked it because it was all they could afford.

But I could afford an Amiga, and had one, yet still used my C64 a lot of the time for coding, for demos and for games, and so did many of the people I went to school with at the time. While a lot of people couldn't afford alternatives, a lot of people also could afford them and either had those alternatives but still also used the C64, or could afford something else but had other priorities and saw the C64 price point as just low enough to make it worthwhile.

By the time I "retired" my C64 in the 90's, I had two heavily expanded Amiga's. First at that point my C64 saw little enough use to be put aside. From what I saw at the time, that was not unusual - a lot of both Amiga and PC users held on to a C64 for years to scratch itches their other computers did not.

When I went to the first The Gathering [1] in '92, while I took one of my Amiga's, there were still plenty of C64's there with people who spent more on peripherals than it would have cost them to buy an Amiga or PC, where it was a clear choice.

[1] https://en.wikipedia.org/wiki/The_Gathering_(LAN_party)#Begi...

The article is right if you look at computers as standalone devices. The problem is, they aren't standalone devices for most people. Even prior to the popularity of the Internet, people exchanged documents. With the rise of the Internet, people started consuming more and from more sources. Interoperability became more important, and other people started defining standards that were well beyond the capabilities of 8-bit micros. The C64, among others, were doomed.

(For what it's worth, I was encountering people who used 8-bit micros well into the late 90's. They were happy with what they had, but they were also very much using their computers in isolation.)

They were cheap toys. That's all. They were usefully educational if you wanted to understand the internals, but 95% of the time they were for entertainment.

Modern computing uses the same hardware principles, expanded and speeded up by many orders of magnitude. But a hugely bigger set of affordances.

Mass storage just works without floppy swapping or cassette saving, data and applications are distributed, there's pocket as well as desktop access, and cameras, sound, and video are a given. And that's just the hardware.

It's like comparing a high speed rail network to a skateboard.

Yet high speed rail networks are practically useless for transportation within one's community (or even within most cities). In those cases, modes of transportation such as walking, skateboarding, cycling, driving are more effective.

Using a C64 or C128 today would be kinda silly, but that wasn't the case in the late 1980's through mid-1990's. They were fine for many smaller tasks, and a great deal better than the alternatives if you could not afford a "real computer". Even as a glorified typewriter, a C64 was orders of magnitude better than an actual typewriter. Even as a calculator, a C64 was order of magnitude better than an actual calculator.

Of course cost is the deciding factor here because computers aren't the same as transportation networks. Aside from cost high end computer is going to be better than a high end one in every case. It has physics working for it, rather than against it. Yet cost of replacement is still a factor.

I mean, he was wrong in many ways.

However, old computing does have a place, as has been mentioned before here on HN, Ben Eater has done an awesome job at getting folks involved with hardware and low-level software design:

YouTube: https://www.youtube.com/c/beneater Reddit: https://www.reddit.com/r/beneater/

Note that I don't know the guy (though I really wish I did, would love to have a couple beers with him.) I stumbled across his stuff just this year and loved it thanks to my background along with my family background.

In the absence of elementary silicon design classes, we find YouTubers stepping up. I am rather critical of such folks, as most tend to be super pushy wish sponsorships, ads, and what not, however, his videos (thus far) do not do that. Note that I did (eventually) buy a couple of his kits.

In terms of compute requirements, 640x480 is quadruple 320x200.

Double would be 448x280.

I find it hard to believe for a C64 (my first computer) but oddly not for the Amiga 500+ (my third)

With that machine my nostalgia is absolutely maxed out :)

Right there with you. My computers in order were:

1. Commodore 64

2. Commodore 128

3. 10 MHz 8088

After the 128 I considered a) Buying a Lt Kerrnal 20 MB hard drive b) Buying an Amiga c) Going IBM Compatible

Thank God I went with the 8088 but I probably own 15 Commodore 64s now. I taught myself component level repair on them so I buy them broken for cheap and repair them. I have a full set of 64 and 128 with matching CRT monitors and Ultimate II+ and more.

I have a similar route. For me I had a tiny Carry-I PC 8088 that I still regret selling to this day.

I have great nostalgia for its twin 3.5” floppy drives, Microsoft mouse (Windows 2) and orange monochrome display.

Man I wish I’d bought an Amiga at the time though. With hindsight PCs were so crap back then, even if the limitations were fun.

I hope that you have a museum in your will, and that they are aware of that.

> If the need continues to exist until the year 2000, then that machine is still satisfactory.

Funnily enough Melbourne Metro used C64 systems well into the 2000s for displaying timetables.

Yeah, it's totally myopic

The C64 wasn't even a good 8-bit machine for a start

And even if it was, it is so removed from today's computers to be useless. Really

I'd say a 386 is not obsolete from that point of view (of the discussion). Or even the Amiga as others mentioned. But the C64 is.

I am curious which 8-bit machines you thought were good, if the c64 wasn’t.

MSX 2 and 2+ for example

Amstrad CPC6128 and Enterprise 128 were arguably much better machines than the C64, architecture-wise. The idea that the C64 was the 'best 8-bit machine of the market' is very US-centric - there were certainly far better 8-bit machines on sale at the time, its just that their market wasn't the USA.

(Disclaimer: Oric-1/Atmos user here. A much better machine than the C654 - it taught me to program.)

I don't think anyone argues that specific c64 hardware designs, like how the video chip steals CPU time, were all that brilliant. The 6510 chip has some strengths and weaknesses compared to the Z80-based micros. Most micros beat the c64 on filled polygon 3d graphics for instance. But the c64 had the best sound by far of the 8 bit micros, even those very late models that you mention.

The MSX, MSX 2, Amstrad CPC and Enterprise all had better hardware and ROM software but they came out later than the C64 (fall 1982), which by then already had lots of games and dominated the market sales-wise. As today, we buy computers and choose operating systems based on the software we want to run. Which would you buy? The system with the most software and the most software currently being made, or a system with somewhat better hardware but little software available?

When the C64 came out, it had the best graphics and sound, but not the best ROM software. I wouldn't even say it had an operating system. It was really just a small, simple KERNAL with some system calls. It didn't have a very good BASIC either, and couldn't auto boot. It was essentially a video game console with a keyboard and a slow floppy drive so parents would think it was a computer. It didn't even have a proper backward compatible upgrade path. The C128 and C65 aren't backward compatible with it, they have a C64 mode.

I had a VIC-20 and then a C64 in the 80's but looking back now, I'd rather have had an Atari 800/XL/XE. The graphics and sound were almost as good but it had a real OS, custom DOSes, and lots of interesting programming languages available for it. But I, as many others in software and IT today who got started with Commodore computers, couldn't afford the Atari or Apple systems so I am grateful Jack Tramiel made computers for the masses, not the classes.

The Amstrad CPC 464, 664, and 6128 didn't have hardware scrolling or sprites, but the 464 Plus and 6128 Plus did, making them more comparable to the Commodore 64 and 128.

The C64 was pretty much the best of all the 8-bit machines of that epoch, except maybe for super high-end MP/M boxes. It shipped with the full complement of RAM such machines could handle without banking, its CPU was a lot faster than that of the CP/M machines and TRS-80s, it included a sophisticated audio synthesizer, and it had the kinds of color and hardware-accelerated scrollable tiles and sprites that made it possible to do decent 2-D games. Its text resolution wasn't anything special but at least it had lowercase, and it had enough RAM to bitmap the whole screen, even if you couldn't put arbitrary colors everywhere.

The NES (a few years later) had better graphics, and the Apple //e had better text, but the C64 was pretty much the overall champ.

There are 8-bit machines today like the Arduino that can run circles around the C64 in some ways, because they're made with dramatically better fabrication technology so they can run much faster; is that what you mean about it not being a good 8-bit machine? I've synthesized audio on an Arduino that you couldn't have done on the C64.

Obsolete, though, it certainly is.

C64 was very similar to the earlier Atari 800, just much cheaper in terms of price and construction, and (reportedly) sold below cost to drive the minor players out of the business. Nobody cross-shopped a CP/M or IBM PC with a Commodore.

Note the article is from the trailing edge, and the other 8-bit systems pretty much almost dead at that point, and the C64 was just barely still selling for another year or so on cheap fumes.

(Never was even aware of C128, so no comment on that.)

The IBM PC was 16-bit, though, like the TI 99/4A and SNES, despite its 8-bit data bus. The C64 was introduced in 01982, and people kept selling CP/M systems and TRS-80s for several years after that. I agree CP/M was pretty dead by 01988.

Commodore seems to have really believed obsolescence wasn't real. WDC launched the 65CE02 in 01988, and Commodore tried to build the Commodore 65 on it in 01991.

A decision I've never understood from that period: IBM introduced the PS/2 Model 30, with an 8-MHz 8086, in 01987. The 8086 was from 01976, but at least it had a 16-bit data bus, and at 8 MHz it was over a MIPS (though only 0.4 Dhrystone MIPS if we believe http://www.netlib.org/performance/html/dhrystone.data.col1.h..., that's still 8 times the Cromemco Z2 with its Z80, and 20 times the C64, though at this point we're well outside the area where Dhrystone is a good measure). They did replace it with the "PS/2 Model 30 286" the next year, but unfortunately not before my school district got suckered into buying some of the crippled beasts.

I agree that the C64 was pretty similar to the Atari 800, but it had much better sound, it had 256 multicolor sprites instead of 8 two-color sprites, and the 800 shipped with 8K and couldn't hold more than 48K. So there was a pretty big difference in what you could do on an Atari and what you could do on a C64.

How far in the future are you from that you start every year with a leading zero?

I suspect some affiliation with https://longnow.org/, they're notorious for 5-digit years.

A minor quibble: The C64 had 8 multicolor sprites. You can display more, but it requires multiplexing them vertically using the raster interrupt.

Thank you for the correction! You're right, it had only 8.

Regarding the PS/2 Model 30 using an 8086, it's important to remember that average consumers didn't quickly adopt the latest microprocessors like they do today, and it was very common for manufacturers to use much older chips in newer products. Hardware was expensive, and the latest hardware was crazy expensive -- the PS/2 Model 30 with its 8086 had a list price of $2,295 ($5,952 in today's dollars), while the contemporary PS/2 Model 80 with its 80386 was $6,995 ($18,142 in today's dollars!). Also, the software commonly in use on IBM compatibles at that time (i.e. DOS) wasn't particularly suited for taking advantage of the newer hardware.

My school had a number of 8086-based PS/2 Model 25's. By the time I used them in 1991, they had been retired to second-tier labs, but they were still perfectly usable for checking email, word processing, telnet to the UNIX server, and some light games. (I remember Star Control ran particularly well on its quasi-VGA "MCGA" graphics.)

To be fair, with Contiki, a Commodore 64 is also usable for checking email, word processing, telnet to a Unix server (if you don't really need 80 columns), and games, even though it's much slower than the Model 30 or Model 25 and has less RAM.

Indeed! I'm familiar with Contiki, and have even done a bit of Contiki programming for fun. :) Sadly, it wasn't around back then, but it's fascinating to see how more value can be squeezed out of old systems.

These particular PS/2 Model 25's probably had an edge on the C64, though. Having proper Ethernet cards and access to Netware servers allowed loading applications a lot faster than, say, a 1541 disk drive. Although my present-day C64 with Ethernet adapter, SD2IEC, etc., might could give it a run for its money. :)

fyi, the 0PS/2 Model 030 was always 0junk, it really only existed as discipline and punishment from 0IBM's 0enterprise 0sales.

C64 production continued until 1994. Commodore announced that year, just prior to its bankruptcy, that the C64 production would finally end in 1995, but then they went bust. They held onto the C64 production that long because it remained profitable for them.

Commodore didn't have the money to sell the C64 below cost. I don't think there's any evidence to suggest they did other than possibly for very short periods. People thought they did at the time, sure, but Commodore's ability to cost-cut was in large part due to their extreme vertical integration at the time.

In 1988 the C64 was still selling over a million units a year. It was losing market share fast, but due to the growth of the overall market, not due to dropping C64 sales - sales at that point were still stable. The NES sold ~7-8m in 1988.

C64 software sales were already dropping off dramatically by 1988, which is what this article is alluding to.

Yeah, maybe it remained the "pirate's choice" in certain circles, as a lot of the adverts there were for copy carts etc.

Does the C64 does this: https://www.youtube.com/watch?v=-6qGP1xP1EA

So, no, it wasn't the best

Those graphics look comparable to the NES, which as I said above had better graphics than the C64, but it seems that the MSX2 had a much slower CPU (a 3.58 MHz Z80), and possibly as a result it apparently didn't have a GUI environment comparable to the C64's GEOS. Also it only had 48K.

I'm confused, because if I look at Wikipedia it lists the CPU for the C64 as "MOS Technology 6510/8500 @ 1.023 MHz", whereas the (newer) C128 has the same 4MHz Z80 CPU that the MSX had. But you're saying the C64 was actually faster?

> Also it only had 48K.

I think you mixed up the numbers here? The 48K refers to the ROM size (C64 had 20K), not RAM (MSX2 machines commonly came with 128K RAM, but you also had 256K models).

The C-128 had both a Z-80 for CP/M, and a 2 Mhz version of the 6510 (the 8502). Could only run the 8502 at full speed with the regular graphics processor off/using the 80 character display.

You're right about the RAM, and I was wrong. Yes, a 3.58 MHz Z80 (or even 4 MHz) is generally slower than a 1.023 MHz 6502, but it's not much slower as I said. As pkroll points out, the C128 also had a 6510 as well as the Z80—Commodore was hardly going to ship a C64 follow-on that couldn't run C64 software!

The C64 didn't have that horrible sound chip, no. :-)

The MSX2 was certainly fascinating, and better on most graphics measures (more colours, more sprites, better scrolling - unless you use the unintended C64 VSP techniques, which breaks on a small number of machines.) However, it had far inferior sound, and despite the higher clock rate it wasn't usually faster, since most 6510 instructions execute in 2-7 clock cycles whereas the z80 typically uses 5-25 cycles. The later MSXes could in theory have a lot more banked memory than the c64, but I don't know how common that was (there were RAM extenders for the c64 too but they were vanishingly rare).

Apart from more colours and higher resolution, there's nothing in those 50 games you linked that would be clearly impossible on a c64. You should link to a demoscene showreel instead ;)

https://www.youtube.com/watch?v=EcVa0zviOJw&list=PLFxDEp8RK4... is a C64 demo playlist, but I'm not sure it shows anything a MSX2 couldn't do either, except for much better sound. It certainly has a lot of stuff that's more visually impressive than those MSX2 games, but I suspect that's because it was made by better programmers and graphicians, not because of the limitations of the hardware.

Definitely if I'd seen stuff like those C64 demos in 01982 when the C64 came out, my reaction would have been total incredulity; it is so far beyond what I imagined computers were capable of that it would have seemed like alien technology. And that is not true of the MSX2 games. But to a significant extent that's not really about the capabilities of the computers themselves but about the SID and the VIC-II, and to a smaller extent it's about algorithms and tricks that were developed years after the C64 itself. Boogie Factor, for example, is from 02005.

MSX* demos I've seen like https://www.youtube.com/watch?v=busiDuS94qY are more visually impressive than the games linked upthread, and very impressive indeed for a Z80, but still not in the league of the C64 demos, despite having lots more RAM to work in. I suspect the anemic CPU is to blame; that demo leans on raster shears pretty hard.

Really, though, I'm more convinced by GEOS. In GEOS the C64 had a WIMP GUI with a WYSIWYG word processor with multiple proportional fonts, a pixel paint program like MacPaint, a faxing program, etc. Because what's really important is how computers increase human ability and especially how they empower human creativity, not how many colors they can show.

This is from 1988, where the incredible upgrade cycle was about to start. Standards for consumer HW went in 4-6 years from 8086s to 486s, from Hercules/CGA to SVGA, from 5 1/4" 360 kb floppies to 400mb harddisks. I think the comment just about predates this incredible upgrade cycle.

I mean, that comment predates the "Andy & Bill law" [1]: what Intel giveth, Microsoft taketh away. There was no need to upgrade stuff back then.

In that context, not a very weird comment - especially for someone whose job it is to sell these machines.

[1]https://en.m.wikipedia.org/wiki/Andy_and_Bill%27s_law

  > There is no such thing as hardware obsolescence.

That's a bit naive. Older hardware devices, such as Android 4 devices, cannot be upgraded to TLS 1.3. Hardware also becomes obsolete when wear items are no longer available - and that could include 3.5" floppies, or compatible batteries, or even RAM modules in some applications. Not to mention wifi devices that support only 802.11a, or G1 cellular phones, or even analogue television in the United States.

Often devices are still needed, but the ecosystem surrounding their usage has changed. That is what often brings around obsolescence in this industry.

> Older hardware devices, such as Android 4 devices, cannot be upgraded to TLS 1.3

That is software obsolescence, not hardware.

His job was to sell 8 but computers so that may have influenced his viewpoint or what he wanted to say on the record.

> There is no such thing as hardware obsolescence

I guess if you assume that computation requires no energy…