#: 8922 S12/OS9/68000 (OSK)
    27-Dec-90  23:54:00
Sb: #8921-#68000 ASM Language
Fm: Jack Crenshaw 72325,1327
To: Jack Crenshaw 72325,1327 (X)

[Continued]

I've identified three main kinds of instructions:

(1)  The binary instructions, like   D0 = X or D0 + D1.  If someone _REALLY_
wants to keep a C-like syntax, I'll settle for D0 += D1, but I sort of like the
stark simplicity of the first form.

(2)  The unary operators, like not ( !D0), negate (-D0), and shifts ( >>D0 or
<< D0, 4 )

(3)  Things like LDIR that are better left as mnemonics.  It's best to think of
these as "built-in function calls."

Whichever approach I take (modified traditional, or C-like), it's important
that the syntax be written so it can be handled by a predictive parser. That
simply means that when you see a certain character (or, more generally, a
token) you always know what to do.  That, in turn, means that constructs like
-(A0), which starts out looking like an expression, are N.G.

Hope this stimulates some thought & discussion.  I'll build the assembler if
(a) Anyone cares, and (b) we can agree on what should be built.

I've gotten one mail message so far that said don't do anything.  That may be
the consensus, but for those who are concerned about compatibility, I'd like to
point out:  A screen editor like Brief, with good global substitution
facilities, can take care of different notations pretty quickly.

Jack



There is 1 Reply.