Re: [colorforth] forth taxonomy

[Mark Slicker <maslicke@xxxxxxxxxxx>]

On Sun, 3 Aug 2003, Roelf Toxopeus wrote:

>
> On Sunday, Aug 3, 2003, at 02:02 Europe/Amsterdam, Mark Slicker wrote:
>
>   I would assume that someone implementing his/her Forth would be
> interested
> in the relation between this Forth and the underlying hardware.

Of course.

>
>   One of the 'hot' implementing technique items at the moment is compiler
> optimization.

This is interesting to a point. Any competitive optimizer is going to add
signifigant complexity. By competitive, I mean to take a set of general
benchmarks and compare results with another compiler.

I'm interested in applications first and foremost. The assumtion is that
the Forth programmer writes good code; this can eliminate a whole catagory
of optimization. I will analyze code for patterns that can be eliminated
or improved. Chuck's peephole optimization will be a basis for this.

My harware is a PDA which uses a StrongARM RISC processor. If I succeed in
creating compelling applications, perhaps there would be economic
incentive to create a Forth on Forth chip PDA.

>   Something else you might find interesting is the study from Reuben
> Thomas
> with respect to a machine- and native Forth implementation on an ARM
> processor:
> http://dec.bournemouth.ac.uk/forth/euro/ef99/thomas99a.pdf
>

I've looked at that, but I don't draw many conclusions. His Machine Forth
in not quite like colorForth.

Thanks for your input,

Mark

---------------------------------------------------------------------
To unsubscribe, e-mail: colorforth-unsubscribe@xxxxxxxxxxxxxxxxxx
For additional commands, e-mail: colorforth-help@xxxxxxxxxxxxxxxxxx
Main web page - http://www.colorforth.com