RE: Why can't I re-run this program?

Hi Ray,

Yes, I've seen some of the math operations. Looks like a complete set.

I like to be able to comment out blocks ('C') of code, a "jump" around
them would be nice, but at least we have the '/' block delete for Gcode.

I seem to remember some time ago that we were talking about a "higher
level" of part program, above the Gcode level. I will look for the
papers on NIST.

TFTC (Thanks For The Comments)

Alan KM6VV


Ray wrote:

>
> From: Alan Marconett KM6VV <KM6VV@...>
>
> >Hi Ray,
> >
> >Thanks for the comments. <s> I guess the "programmer" in
> >me wants to think about reusable code, modularity, ease of
> >understanding, etc.
>
> Don't we all! If you chase down the roots of G-code variation in
> commercial machines it comes down to sales posturing. "With the
> competition's machine you will have to write 20 blocks of code. With ours
> you can do it in one!" <g> There have been positive results to the
> language from some of this but it does tend toward confusion.
>
> If you chase down the roots of the "hobby" grade controllers, most started
> with bright folk who had a dream and wrote some code. Most of these were
> less interested in rigorous interpretations of RS-274 motion control
> language than they were in making their software and hardware work. The
> result is everything from HPGL -> dxf -> a g-code flavor -> conversational.
>
> >I was also wondering about conditional statements; if we have labels for
> >subroutines, do we have something in NGC or another "flavor" Gcode? We
> >could use labels for "jumps", i.e., "#IF V200 = 0 GOTO P20", Ok, the
> >syntax is not there yet (make it look a little like basic?). The V200
> >would be the variable, and P20 would signify the point in the program
> >that is labeled $20. Possibly "G23 #200 = 1.200 P20"?
>
> There are conditional (boolian) statements available in the NGC interpreter
> but they aren't be used for jumps or loops. The line numbers are not used
> for anything except reference to humans. Variables are subject to math
> and can be applied to distances and codes but again there is no good way
> within the interpreter itself to use them for jumps.
>
> IMHO - NIST has taken g-code interpretation about as far as it wants and
> needs with the current six axis system. For the future they seem to be
> looking to other approaches to translate geometry or real or drawn objects
> into machine motion. (There are a bunch of interesting papers on their
> site)
>
> HTH
>
> Ray