- Quite bold to put mit license and claim copyright/authorship on Claude generated code.
- Fair point on transparency. I've updated the README to credit Claude Code for the assembly. I directed the architecture, feature set, and all version decisions - Claude wrote the code under that direction. Seven releases over about 6 weeks, commit history and changelogs document the progression. Happy to discuss the design decisions if you're curious.
- How do you know? The excessive comments?
- Author is all over AI, other projects on his github are some AI nonsense made with Claude and at least partially somewhat credited this way (or maybe not obscured well enough with Claude signed commits). Release page is your standard affair LLM emojis gag. But mostly code itself smells llm, was dumped all at once and there are no commits since initial upload.
- Line 2948: jsr update_day_rollover ; <-- ADD THIS
:-D (etc. etc.)
- Project time/resource constraints not withstanding, wonder how this compares to other c-64 os alternatives.[0]
?? future options ?? : Unix bootable from a cassette tape drive? :-)
?? c-64 unix server with ports to hook up one or more c-64's via CaTer (Cartridge Terminal) -- perhaps best done via tcp-ip 'ports'.
?? fuse[1] for c-64 to be able to 'read' other floppies / c-64 unix swap space over tcp-ip.??
?? "modern" 5.25" for c-64 via c-64 variant of greaseweazle[2] ??
-----
[0] "The Many Operating Systems of the Commodore 64" : https://lunduke.substack.com/p/the-many-operating-systems-of...
[1] c64-fuse (perhaps over c-64 tcp-ip) : https://en.wikipedia.org/wiki/Filesystem_in_Userspace
[2] greaseweazle : https://github.com/keirf/greaseweazle
- Around 1987 I mostly completed a Unix-like OS for the C-64 called MATRIX. I was probably around six weeks away from burning it to a PROM when I got a new girlfriend and completely lost interest in the project.
I don't remember too much about it, other than:
- Because Commodore drives had ludicrously long file names for the era, paths like /etc/dev/joy1 didn't need any weirdness.
- Password encryption? What's that?
- What we would call "metadata" today was stored in USR files.
- Directory listing was agonizingly slow. I remember commandeering tracks 16 and 17 for my own hair-brained directory structure in an effort to speed things up.
- This is a great story, and you're further along than I ever got in 1987! I had a C64 back then, too, and was fascinated by it, but never attempted anything this ambitious. Girlfriends, too, got the best of me! Fast forward nearly 40 years, and I finally built my Unix-inspired shell for it, just with a very different kind of assistant helping with the assembly. :)
The directory speed problem is real. I sidestepped it entirely by keeping the filesystem RAM-resident (max 8 entries, heap at $6000), which makes LS instant but obviously volatile. Your track 16/17 commandeering approach is incredible and fascinating. MATRIX sounds amazing, and you should dig it up and finish it now! :)
