4 ms·

An FPGA implementation is missing the big educational aspect though, namely that this product probably intends for it to be feasible (just like it was with the

by anyfoo 5y ago

An FPGA implementation is missing the big educational aspect though, namely that this product probably intends for it to be feasible (just like it was with the 8 bit computers back then) to fully understand the hardware, prod it, modify it, and interface to it. Usually up to IC boundary, but the key there is that the ICs are distinct and (hopefully) not too complex. Even a fully "open" modern platform is usually just way too complex, integrated, and often full of abstraction layers. With an FPGA, you in theory have that within the FPGA design itself for the most part, but it's still integrated and wrapped away so much that the directly relationship between current and logic, and things like an actual external bus with traces you can probe and hook into, is just not as tangible.

vidarh 5y ago
To be clear when I wrote about my preferences above it was from a perspective of what I'd like to have and to use myself. I absolutely agree with you that many of these are great for educational purposes, so I absolutely think it's great some people are doing these too. That said, for that purpose I'd have stuck with something a more direct analog to one of the existing 8-bit machines myself. But it's not my projects so he is of course entirely free to make it however he prefers.
cmrdporcupine 5y ago
I personally didn't find Verilog all that hard to learn, and it's frankly simpler to understand and modify the "hardware" in that form than using a multimeter and logic analyzer and futzing with solder. I built myself a simple retro-style machine around a RISC-V core and my own custom designed video controller in a few weeks, without any prior Verilog experience. Although I am a software engineer, I don't feel like those skills actually helped that much.
- crispyambulance 5y ago
  Kudos to you if you can go from zero to your own custom video controller implemented in an fpga in a few weeks. But for normal humans that just want to get involved with hardware and dabble somewhat with software/firmware in the context of vintage computing, taking a detour into fpga-world would be a deep unpleasant rabbit hole unless one was specifically interested in fpga's and there was no other viable option (which in the case of custom ASIC's is admittedly quite possible). Compared to the nightmarish complex toolchains used for fpga's, digging up some datasheets and working with a DMM, scope, and logic analyzer is a breeze and actually quite pleasurable.

by anyfoo 5y ago

vidarh 5y ago

To be clear when I wrote about my preferences above it was from a perspective of what I'd like to have and to use myself. I absolutely agree with you that many of these are great for educational purposes, so I absolutely think it's great some people are doing these too. That said, for that purpose I'd have stuck with something a more direct analog to one of the existing 8-bit machines myself. But it's not my projects so he is of course entirely free to make it however he prefers.

cmrdporcupine 5y ago

I personally didn't find Verilog all that hard to learn, and it's frankly simpler to understand and modify the "hardware" in that form than using a multimeter and logic analyzer and futzing with solder. I built myself a simple retro-style machine around a RISC-V core and my own custom designed video controller in a few weeks, without any prior Verilog experience. Although I am a software engineer, I don't feel like those skills actually helped that much.

crispyambulance 5y ago
Kudos to you if you can go from zero to your own custom video controller implemented in an fpga in a few weeks. But for normal humans that just want to get involved with hardware and dabble somewhat with software/firmware in the context of vintage computing, taking a detour into fpga-world would be a deep unpleasant rabbit hole unless one was specifically interested in fpga's and there was no other viable option (which in the case of custom ASIC's is admittedly quite possible). Compared to the nightmarish complex toolchains used for fpga's, digging up some datasheets and working with a DMM, scope, and logic analyzer is a breeze and actually quite pleasurable.