3 ms·

I remember an episode of Computer Chronicles where a man claimed 8-bit computers should be as relevant today as they were when just released, because "every pro

I remember an episode of Computer Chronicles where a man claimed 8-bit computers should be as relevant today as they were when just released, because "every program is a turing machine, essentially." His main argument was that the memory and CPU limitations helped curtail software feature bloat. When asked about the possibility of advanced Multimedia features becoming more desired among consumers, he shrugged it off, saying an expansion peripheral would do the job just as well as buying a new machine, without having to learn a new machine entirely. Now that we have cartridges for the C64 that allow for massive data storage and 32+ Megs of RAM, maybe he was right. We could have all been using C64s or Amigas with a ton of expansion cards/carts installed.

stordoff 7y ago
> His main argument was that the memory and CPU limitations helped curtail software feature bloat I think there's some truth to this - if you have a limited amount of RAM, you can't just throw RAM away and do the easy thing. You have to be careful with it. The flip side of that is of course one man's feature bloat is another man's valuable feature, and some things just _won't_ fit in the limited amount of RAM.
- sneakernets 7y ago
  Seems to be the "Usable, Extendable, Efficient - pick two" scenario at its core. I remember very well a word processor on C64 (GeOS Writer I think?) that would use its own disk as swap to get around memory limits. I eventually bought a memory expansion just to get around this, which wasn't cheap in the slightest. And that's just for text! It's stunning to me how inefficient text can be when every byte counts.
mntmoss 7y ago
You get bottlenecked on I/O eventually, and that is the factor that creates jumps to new platforms. Even with modern devices, I/O is a huge point of contention: we argue for our serial ports and 3.5mm headphone jacks, but we upgrade for the newest USB or DisplayPort iteration. When I started viewing computing in terms of connectivity, the device categories and developments made a lot more sense. We were fine with switches and printouts - except that we wanted to input and retrieve things faster, so we ended up with display, keyboard and mouse, storage media and networking, etc. The Turing Machine parts are important, for sure, but they are most important in pushing the I/O to capacity, and only sometimes get to exercise abstract computing prowess.

stordoff 7y ago

> His main argument was that the memory and CPU limitations helped curtail software feature bloat I think there's some truth to this - if you have a limited amount of RAM, you can't just throw RAM away and do the easy thing. You have to be careful with it. The flip side of that is of course one man's feature bloat is another man's valuable feature, and some things just _won't_ fit in the limited amount of RAM.

sneakernets 7y ago
Seems to be the "Usable, Extendable, Efficient - pick two" scenario at its core. I remember very well a word processor on C64 (GeOS Writer I think?) that would use its own disk as swap to get around memory limits. I eventually bought a memory expansion just to get around this, which wasn't cheap in the slightest. And that's just for text! It's stunning to me how inefficient text can be when every byte counts.

mntmoss 7y ago

You get bottlenecked on I/O eventually, and that is the factor that creates jumps to new platforms. Even with modern devices, I/O is a huge point of contention: we argue for our serial ports and 3.5mm headphone jacks, but we upgrade for the newest USB or DisplayPort iteration. When I started viewing computing in terms of connectivity, the device categories and developments made a lot more sense. We were fine with switches and printouts - except that we wanted to input and retrieve things faster, so we ended up with display, keyboard and mouse, storage media and networking, etc. The Turing Machine parts are important, for sure, but they are most important in pushing the I/O to capacity, and only sometimes get to exercise abstract computing prowess.