Notices by Vertigo (vertigo@mastodon.social), page 16

When your code looks like a back-slash, that's a unit tester's worst nightmare.

@elomatreb @jk Disclaimer: I don't actually follow Von Neumann's proofs. It's well above my head. But I've read about them before and found the idea interesting.

@elomatreb @jk The thing is, it's not wrong when you work in Von Neumann numeral system. You can map each '1' to a set of empty sets (what Von Neumann used), so If 1 = {}, and 11 is {{}}{}, and 111 is {{{}}{}}{{}}{}, etc. then it's provable that 2 is greater than 1, and from that, the fundamentals of basic arithmetic can be logically proven.

If only you'd known about Von Neumann's numbering system and attempts to prove 2 > 1, then they might have advanced you a few grades instead.

@zuz Depending on my calendar, I'd be down for hanging out at a hack-day as well. Maybe I can use it to work on the Kestrel-3's OS, or perhaps start working on the code generator back-end for my Forth compiler.

@slipstream Sunday.

I .... kind of forgot today was Monday. Whoops.

@gled Thank you. This will be very useful, I think.

@Shamar That's at a skill level beyond me at the moment. I program FPGAs, but don't (yet) build boards for them. I definitely want to build a custom motherboard for my project some day.

Right now, my Kestrel-2DX targets a Digilent Nexys-2 FPGA board, but that's no longer in production. I'll need to port it to something which still is. I'm thinking the Altera DE-1. It's relatively cheap at $150. It'll also give the computer 512KB of RAM (a big improvement from the 48KB it currently has!).

@Shamar I guess 9front is a kind of distribution then, a la 4BSD was back in the day.

@Shamar Nice, I had no idea it also included the kernel. I'll definitely research it further!

@Shamar Not just free software, but also hardware as well. My project is a computer intended to be free of as many of the constraints we find on machines today as possible.

Currently, I have a computer which is basically a 64-bit Jupiter ACE, with a tiny amount of RAM, running Forth. But, longer term out, the next phase of the project is to evolve the design to be more Atari ST-like -- 1MB to 2MB of RAM, 640x480 256-color display, consistent expansion protocols for storage devices, etc.

@Shamar I truly wish I could figure out how to port something the size of Plan 9 myself though. I seem to excel at starting new projects, but when it comes to porting existing projects, I just can't figure anything out. Kudos to you for being able to make the port.

Like, I have the GCC toolchain for RISC-V working; where would I even *begin* to port the kernel? And, how would I even begin testing it to see if it even works?

@Shamar > Had Harvey asked that I wouldn't have contributed a line.

Precisely why I appreciate FSF and GNU project making it very clear up-front, and why I haven't contributed anything beyond bug reports to them. :)

@Shamar
Re: 9front -- I was under the impression that it was just the Plan 9 user-space; that it wasn't the kernel. Am I mistaken?

Re: Free vs Open Source -- I describe my project as FOSS, and license everything under MPLv2. I'm happy to include contributors names in a CONTRIBUTORS.md file, along with a summary of their contributions. I also encourage said contributors to update that file if they think I missed something. I want to give authors the credits they deserve.

@Shamar Re: Harvey -- I definitely feel for the author here, and I'd be pretty miffed too. However, I note that the FSF makes it clear that, if you want to contribute to GNU projects, you must first sign a copyright waiver before your code can land. E.g., after committing, your code belongs to the FSF, not to you.

I guess Harvey's mistake here is not making this plain up-front.

@Shamar Yeah; but it's actually higher level than what 9P exposes. CBM DOS was a true DOS in the sense of CP/M. You could tell it to copy files from drive to drive (for devices equipped with multiple drives, for example; hence the 0 in S0:blah).

One of the neatest tricks is to print a file by telling the disk drive to send a file directly to the printer, then disconnecting the computer from the serial bus.