(Maybe you're looking for my page about Lua and Forth? As of sep/2005, the rest of this page is several years old... Sorry!)
(Update: the PForth implementation part is sleeping while the rest of the project mutates into the Crims; see above.)
The goal of this project is to have a Forth system for Unix that is powerful, free, and fun to learn and to use. "Fun to learn" implies (or almost) that by learning it one also learns lots of useful things, not only the intrincacies of C, assembly and of the design of one particular Forth; "powerful" implies that it should have many of the features found in other languages or programs.
I think that the best way to accomplish this is to write just the code needed to "glue" certain other packages. This will be quick, as there will be little new code to write; relatively bug-free, as most of the code we'll be using was already debugged by their authors; interesting, because it will unveil many tricks about coding and debugging under Unix, and of the low-level interfaces of some programs -- things are often bragged to be easy, but that we all know that are not. And we may even get some support from the people who wrote the programs we'll be glueing...
In the spirit of Project eev, I'm documenting most of the steps taken to reach the present state of the project. The documentation starts very messy, of course, because it is just a record of the steps I took, written in more or less executable language, but it can easily be cleaned up if someone asks to. And the code is very small, so it should be really easy to understand: I'm using mktclapp to glue PForth and Tcl; I'm packing mktclapp for Debian, and I tried to add to the package exactly the docs (for example, this README) that would have made me start playing with it quicker.
At this moment I'm working only on the Tcl<->PForth connection. Right now Tcl can call PForth, and PForth can return numbers and strings, but PForth cannot yet execute Tcl code. A medium-term project is to link Tcl and PForth with Icon, that was my favorite language for many years; longer-term projects include linking to GDB and to TeX, and, in directions orthogonal to these, I'm planning to:
One further note: I'm working on this with an extremely tight budget, and only on my (scarce) spare time... so I'm having to make things fit my current hardware and OS: a pentium Debian GNU Linux box, that will soon also run Debian GNU Hurd. I don't have the resources, in any sense, to play with Windows or other non-free/non-shareable stuff, and also don't have enough practice in dosemu and FPC to be able to play seriously with the existing F21 emulators (JFox's). Also, I can't play with hardware, I don't have any sort of electronic lab, etc, etc, and I know that people only very rarely engage on other people's projects by the net (hey, you! Change that! :-)... but I'm fascinated with Jfox's/Chuck Moore's ideas, and I've had a very good time playing with HS-Forth at the end of my MS-DOS days. Well, let's see.
The code for the Tcl<->PForth connection: at this moment it consists of tclpforth.c, tclpforth.tcl, mtamacros.h and this makefile, all called from this e-script; we also need tclpf_custom.c for the PForth primitives that return values to Tcl, and tclpf_hook.c to build the temporary PForth standalone that is used to generate the dictionary with the new primitives.
See also the Minotaur page for a project that was somewhat related to mine. It seems to be dead.
Help!: people on the net are generally either clickers or people trained in conventional languages, so they tend to measure the value of a program by the number of lines of code it has ("wow, this guy should have suffered a lot to write this, he's certainly a serious professional"), so the eev project have had very little acceptance at the main divulgation channels, and this glue-Forth thing will certainly have infinitely less, no matter how far I go. So, please, if you like some of my stuff give me a link on your page, or, better yet, join, or get in touch.