@davidgro That was a different thing. You can write or read from the physical memory address 0 even on the systems of the today - except that practically all the OS-es are using their paging or memory management to make that address unreeachable. Actually, as far I can remember, on Linux, the addresses between 0 - 65535 are mapped unreachable. If try to write them, you will get SIGSEGV. This is because a memory page is 4K, that is the granularity of the memory permissions. Beside that, we also need SIGSEGV (which would map to a "null pointer dereference" or analog) if it actually tried to access the part of a larger struct (what would be visible as an access to a non-zero, but very small memory address).

C64 had no memory protection, you could read or write anywhere where you wanted. Although it had memory-mapped I/O, meaning that instead of bus-level port communication, all the IO, i.e. communication of the CPU with the chips of the mainboard, happened as if they would be memory. Some chip happily crashed the system on such writes, particularly the CIA did it (general purpose I/O chip and hw clock) for not enough well thought memory operations in the $DC00 - $DFFF range.

You could happily write into 0 on C64. Actually, on the address 1, you could access the chip select pin of some RAM chips, meaning that you could with that switch from the basic ROM into an additional RAM area. Writing to the 0 byte did not had any on the spot visible effect on my experiments, but today I read in the docs, it had actually effect to the GPIO ports related to the memory chips managed at the address 1.

Actually, c64 went even more: the first page, i.e. the addresses between 0-255, very accessible a bit faster, because they required a lesser amount of CPU and data bus cycles. Very important registers, used for important operations, were there.

However, this all is about the C64 hardware. The C64 basic v2.0, which was written by the microsoft, quickly evaporating all my youth nostalgy, had nothing similar to the NULL (of C) or null (of Java). It not even had an "unset", "undefined" or similar type. The 3 types it had, were: 32-bit floating point, 16-bit integer and 0-254 byte long string.

@davidgro The way that #Rust guards memory is completely different to the way Java or Basic does. Rust is a fully compiled language suitable for systems programming, and the memory protection all happens at compile time; the runtime actually gains speed through the compilation process, rather than having the memory protection slow it down.

The real difference is that the Rust compiler actually reads into the code it is compiling, rather than just making a mechanical translation.

That's a question of API, not type system.

It's only enforced because of Rust's strict type system. Python, on the other hand, lets you do whatever you want by comparison, and complains only at runtime. I've seen far too many **kwargs for my liking.

And FD types (e.g. OwnedFd, BorrowedFd) are already in std.

My example would be a thin wrapper around these, most likely. It's only an example of what I'm trying to convey, though.