Ed Minchau wrote: ↑Mon Jul 08, 2024 8:11 pm If we're only using the 6502 commands then there is no point in using the 65816 at all. It isn't just the 32 Rockwell commands we can't use because they won't work on a 65816. We can't use 57 commands that are unique to the 65816 either, because they won't work on a machine with a 65c02. We wouldn't be using the full power of either processor. So why bother?
First, it's not just 6502 opcodes ... it's all 65C02 opcodes
except for the Rockwell extensions to the 65C02 opcodes. JMP (a,X), PHX/PLX/PHY/PLY, BIT #n, etc.
Second, if you are doing something with a VM, you can just "do both". A p-code Pascal could have a 65C02 and a 65816 p-code interpreter. A Forth could have an "original" 65C02 ["o65C02"] kernel and a 65816 kernel. A Sweet16 interpreter could have an o65C02 version and a 65816 version. etc.
________________
encw wrote: ↑Wed Jul 10, 2024 8:27 am Don't worry if it doesn't run on other people's machines?? The problem is that we were promised more than just a hobby machine. ...
But if it is more than just a hobby machine, that means that it continues to also be a hobby machine, and if you wish, you are perfectly free to use it that way. The fact that software that works within the development guidelines may have a bigger prospective install base available to someone only matters for a project if it matters to the project developer. If the project is a hobby project, they don't have to worry about the various pulls that are felt from different parts of the prospective audience, they can just develop for themselves "and anybody else who is interested in something like this".