First Loop


Correspondence leading to first loop challenge ...

Tom -- Let me describe some very small examples that hint at how I would like to program in my particular application. David understands that I am mostly doing this just to see if I can. Let me imagine putting a few instructions into a design database and then generating an absolute memory image from that database in a later pass.

( loop code now on sister site )

My challenge to David is to write an assembler that allows this sort of program manipulation. He has already written an assembler for the Atmel AVR parts that I am using. I can imagine emit using his assembler to produce the final program image. What I'm looking for is tips for representing the program in the tree form that can be inspected (with cycles) and manipulated (with factor-port-setup). My preference would be to do manipulations in a functional style disregarding wasted space.

I did something like this in Java recently where each instruction was a new object. It felt quite powerful. I could run fragments to test them before I assembled them into whole programs. I think forth will be simpler, faster and easier, but we do have to get this design-database representation right.

What is your reaction to this scheme?

Thanks and best regards -- Ward


Last edited April 8, 2008
Return to WelcomeVisitors