3 Jan
2012
3 Jan
'12
11:02 p.m.
On Tue, Jan 3, 2012 at 3:39 PM, Tony Duell <ard at p850ug1.demon.co.uk> wrote:
Sure. Most BASICs had some way to call machine language routines. But
doing so was totally differnet to using built-in fucntions (or using
BASIC subroutines) and requried you gerneally to remember th rright
addresses. Better tnan nothing, but a kludge.
I wrote at lot of hybrid BASIC/6502-machine-code programs back in the
day. In the Commodore world (PET BASIC all the way through the
C-128), you had the USR() function, which was handy if you wanted to
pass one floating-point arg to your program and/or wanted your routine
to pass you back a value - just POKE the address of your code into the
documented USR() vector location so BASIC could hit it via a
JMP-indirect instruction ($6C), or if you didn't need to pass args
in/out (or you were willing to play parser tricks), just directly call
your routine with SYS. It was all clearly described in the manuals -
I wouldn't call it a kludge, just a method that was probably present
in Microsoft BASICs that might not have been present in versions of
BASIC that was, for lack of a better term, more Dartmouth-like.
I didn't ever have much experience in the past with HP BASIC or DEC
EDUsystem BASIC, so I can't guess what, if any, mechanisms there were
for those commonly-encountered-at-school environments.
I guess if you were writing your own machine code, as I was doing,
you'd know your entry points, vs trying to make use of someone else's
code, where I can see how it might seem arbitrary and kludgey,
especially if they didn't prepend things with a jump table.
-ethan
In genral, BASICs were not extendable like that. It is
very difficult to
add new keywords ot a BASIC interpretter, and the way to do it is not
often docuemnted.
MICROS~1 BASIC typically has a somewhat awkward DEFUSR function in the
language to let you create and call machine language code...