I think I never mentioned this before, but there is a simple trick that uses an unconditional label interceptor to detect whether something is a name of an existing instruction: Because the interceptor is unconditional, it is executed even inside the "if 0" block, but only when the initial symbol of the line is considered a label, which implies that it is not recognized as an existing instruction.

I'm posting this as I'm designing a new set of CALM-specific operations that may allow to perform checks of this kind in an "official" and reliable way. Still, I think it's an interesting trick that demonstrates how some of fasmg's cogs intertwine.

This example also makes clear an interesting distinction of fasmg:

From a HLL perspective instructions are often thought about as larger puzzle pieces - containing their parameter syntax. Whereas in fasmg a symbol can pass the instruction test and later fail to match parameter requirements during use.

For example, the DQ instruction cannot be used on a line by itself - that is invalid. This does not cause DQ to be interpreted as not an instruction, but an invalid use of a defined instruction.

DQ can be used as not an instruction by prefixing it with a question mark - ?DQ will be seen as a label.