https://github.com/siraben/zkeme80



25 days ago I started work on a new, Scheme-based Z80 assembler to relieve the problems of buggy macro systems in other assemblers. Soon it spun into a project to create the next extensible operating system for the TI-84+, one that was based on the programming language Forth. And here it is, I call it zkeme80, pronounced like "scheme 80" but with a z sound. It's not done yet (still missing a REPL) but the core is mature. The menu you see in the demo was written entirely in Forth, and you can imagine people swapping Forth programs with each other, customizing the boot sequence to their preferences.

Another important design goal was for this assembler to just work, portably. That's why I wrote it in Guile. The only command you need to run to build the operating system is
Code:


It takes 13.5 seconds (real time) to compile on my MacBook Pro 11,1 running NixOS and 0.5 seconds (real time) to recompile, assuming only changes in the Forth layer.

Any and all help would be greatly appreciated!

siraben

Finally; a well designed OS I feel like this should be getting more attention, but alas everyone seems to have transitioned to the CE. I did clone this and it built just fine on Linux; looks like I'll have to dig out my TI-84+ to give this a try

Also I wasn't quite sure what this means? :



Anyway, sweet project I may have to try out creating some Forth programs; I haven't used that language in forever...

Thanks for the reply . Yeah, I wish more effort is put into the TI-84+, it's a really good device.



The assembler is written in Scheme. So you don't write


Code:


Instead you write


Code:


See the various .scm files in my repository for more.

I personally love the little touches from the screenshots such as the icon highlighting and the loading/shutdown screen animation etc.

I don't know a great deal about Forth, but do you foresee users being able to port code across easily enough?

Thank you! Any respectable operating system should have some sort of user friendly interaction to hide all that complexity

The idea is to have a somewhat fixed kernel/core system and allow people to program the operating system in Forth. I foresee users being able to port existing Forth code. It would be possible to port this to other calculators too, with the necessary modifications to the boot process made. I chose to define words in Z80 for speed, but if need be, most of the language could be defined in itself, so that to port the entire interpreter and compiler only a new primitives need to be rewritten.

I made sure from the start that a significant portion of this Forth implementation is ANS standard conforming, so that code like https://en.wikipedia.org/wiki/Forth_(programming_language)#A_complete_RC4_cipher_program works in zkeme80 without any modification, for instance. I also wrote a new assembler so that building this operating system can be done on any platform with a Guile implementation. I also working on an ASM -> Scheme parser so existing ASM code can be integrated easily.

Hey I noticed this was featured on OSnews.com https://www.osnews.com/story/129236/zkeme80-a-forth-based-os-for-the-ti-84-calculator/

And HaD: https://hackaday.com/2019/01/25/pocket-forth-invades-your-ti-calculator/

Nice .