I think you missed the point. Imagine 2 devices, device A has a chip with flash memory that contains a binary blob with firmware. Device B doesn’t have built-in flash storage so it requires the driver to load the same binary blob during boot. Both devices are reprogrammable and both contain the same closed source firmware. However device A would be allowed but device B would not. From my point of view they are the same device. The fact that you don’t know how to reprogram device A doesn’t make it more or less proprietary.

Why mips and not RiscV? I would assume it’s easier to emulate in software and has good support in linux

EDIT: found it

Some architectures had arbitrarily-shifted operands all the time (ARM), some have shitty addressing modes necessitating that they would be slow (RISCV), some would need more than 4KB to even decode instructions (x86), and some were just too complex to emulate in so little space (PPC).

Could it be the pc relative addressing often used on risc-v would be slow to run on 4004?

I can’t quite understand what is your point? Are you arguing that both JVM and WASM are bad? With this I agree, they both have terrible performance and in an ideal world we wouldn’t use any of them.

Are you arguing that JVM bytecode is better than WASM? That’s objectively not true. One example is a function pointer in C. To compile it to JVM bytecode you would need to convert it to the virtual call using some very roundabout way. But in WASM you have native support for function pointers, which gives much better flexibility when compiling other languages.

Have you seen what it outputs? The same way we can compile C to brainfuck, it doesn’t mean it’s good or is useful.

You can’t compile C to java bytecode, they are fundamentally incompatible. But you can compile C to wasm, which is what you want for a good universal bytecode. Java is shit.

classic

Some people have weird fetishes too, I won’t judge.