I literally came back to this channel 2 days ago and watched nearly all of your videos. I was stunned to see you didn't quit youtube and came back with a huge blast. Keep it up!
Very cool. I've been thinking of similar for many months, now, but don't have nearly the experience you do. This is very impressive. My idea was a sort of a pseudo-assembly-language that could essentially be converted directly to binary instructions to any target, treating them all as, basically, RISC. Since, it seems most procs I've encountered have nearly identical core instructions, and core registers. If that turned-out feasible, then a very limited C compiler was the next step... and eventually an OS. The big hurdles, I guess, is the differing factors like boot-address and memory layout, which would have to be handled specifically for each target. But, if the limited instruction-set/register-set thing works, at least that wouldn't mean learning new mnemonics for each machine. Ideally, that would produce raw machine instructions. And once some underlying "BIOS" was implemented, hardware-specific things like memory-allocation and I/O could be standardized. But then it gets complicated, what about Harvard architecture systems? What about loading several programs simultaneously on systems that don't have, e.g. relative-jumps? Realtime byte-code processing came to mind, but I thought it would surely be too slow to be useful. But your work, here definitely suggests otherwise. Am very intrigued. I suppose it actually makes some sense it wouldn't be much slower than all the BIOS calls necessary... Great work, and got me thinking, thanks!
Amazing work! Beautiful work to see an actual effort to create what could truly be the most platform-agnostic OS in the world! First time I've seen anyone play a game on a 16x2 LCD let alone one as complex as yours. I also wanted to say thanks directly, since 11 years ago not long after high school, you inspired me to build my own TTL CPU from scratch on protoboards. (Used an eeprom with a counter for PC, 4 latches for internal registers, and followed your lead on an addition and subtraction ALU. Had a blast doing it and discovered the fun of creating your own assembly language for myself. Since you mentioned TTL, I imagine it might be possible to even adapt my CPU to run your WheatSystem. I imagine I would basically just need to write a compiler for your kernel, correct? What would be the minimum requirements for an adept style system (registers, ALU functions, What else?), and can the compiler be written and tested in stages? Never did develop mine beyond counting numbers and printing simple messages to the 16x2 LCD, but this OS of yours might be my motivation to play with it again. I have been doing pro C# development since then, but always appreciate literally handling the internals of a CPU, Not something many can say. Thanks for helping set me on to my career in computers.
That's awesome! Have you posted anything online about your TTL CPU? Yes, if you wrote a (GNU-style) C compiler, then you could port WheatSystem to run on your machine. I haven't worked out how much memory is required for a simpler instruction set than AVR, but I'm praying that it's under 64 KB.
YES! Ostracod remembered the password to his TH-cam account! Oh, and just saying, I think it'd be hella cool if Wheatsystem got ported to the DUO Adept.
This is awesome. Love the idea of a high-level environment to work on multiple architectures, reminds me of Java. I was recently thinking of doing a similar idea (albeit MUCH less advanced), where I'd create an emulation layer running natively on a simple-as-possible Harvard 4-bit CPU with large program memory space (about 64KB to 128KB) to emulate a von Neumann 8-bit CPU architecture. The reason for a 4-bit architecture was to minimise on TTL chips as much as possible, with still the possibility of running 8-bit CPU programs. The emulation layer would chew up TONS of CPU time tho, I think the ratio of native 8-bit CPU to emulation 8-bit CPU instruction speeds was about 10:1 lol, making CPU-heavy programs extremely slow, but still theoretically doable.
This is a really cool project, I've been thinking of writing similar things. Did you explore any existing bytecode formats before you started on your own? I think WASM is a pretty good fit for such an OS, or maybe even some mini version of the JVM. The benefit of course being a massive pre-existing tool ecosystem (LLVM can produce WASM for example)
This is a cool project! I’ve been working on a 16 bit computer, kind of like the Amiga, using a 65c816, and I want to try porting WheatSystem to it, at least as an app running under it’s OS.
BreadByteCode is sort of the same idea as "Machine Interface" of IBM's AS/400 computers. From small to really large all the AS400 machine run an emulator which runs the client programs and the OS. (as best I understand it)
I literally came back to this channel 2 days ago and watched nearly all of your videos. I was stunned to see you didn't quit youtube and came back with a huge blast. Keep it up!
Excellent video. This is a great idea. I can’t wait to see where this goes.
Very cool. I've been thinking of similar for many months, now, but don't have nearly the experience you do. This is very impressive.
My idea was a sort of a pseudo-assembly-language that could essentially be converted directly to binary instructions to any target, treating them all as, basically, RISC. Since, it seems most procs I've encountered have nearly identical core instructions, and core registers. If that turned-out feasible, then a very limited C compiler was the next step... and eventually an OS.
The big hurdles, I guess, is the differing factors like boot-address and memory layout, which would have to be handled specifically for each target. But, if the limited instruction-set/register-set thing works, at least that wouldn't mean learning new mnemonics for each machine.
Ideally, that would produce raw machine instructions. And once some underlying "BIOS" was implemented, hardware-specific things like memory-allocation and I/O could be standardized. But then it gets complicated, what about Harvard architecture systems? What about loading several programs simultaneously on systems that don't have, e.g. relative-jumps?
Realtime byte-code processing came to mind, but I thought it would surely be too slow to be useful. But your work, here definitely suggests otherwise. Am very intrigued. I suppose it actually makes some sense it wouldn't be much slower than all the BIOS calls necessary...
Great work, and got me thinking, thanks!
I thought i had subscribe to your channel and aparently had not, took me a while but finally found it! Subscribed, love you work!
This is fascinating! You've earned a sub. I'm excited for more stuff like this! This is an epic way to come back to TH-cam.
Amazing work! Beautiful work to see an actual effort to create what could truly be the most platform-agnostic OS in the world! First time I've seen anyone play a game on a 16x2 LCD let alone one as complex as yours.
I also wanted to say thanks directly, since 11 years ago not long after high school, you inspired me to build my own TTL CPU from scratch on protoboards. (Used an eeprom with a counter for PC, 4 latches for internal registers, and followed your lead on an addition and subtraction ALU. Had a blast doing it and discovered the fun of creating your own assembly language for myself.
Since you mentioned TTL, I imagine it might be possible to even adapt my CPU to run your WheatSystem. I imagine I would basically just need to write a compiler for your kernel, correct? What would be the minimum requirements for an adept style system (registers, ALU functions, What else?), and can the compiler be written and tested in stages? Never did develop mine beyond counting numbers and printing simple messages to the 16x2 LCD, but this OS of yours might be my motivation to play with it again. I have been doing pro C# development since then, but always appreciate literally handling the internals of a CPU, Not something many can say.
Thanks for helping set me on to my career in computers.
That's awesome! Have you posted anything online about your TTL CPU? Yes, if you wrote a (GNU-style) C compiler, then you could port WheatSystem to run on your machine. I haven't worked out how much memory is required for a simpler instruction set than AVR, but I'm praying that it's under 64 KB.
@@Esperantanaso TTL is for n00bs; Imma design an ECL processor!
YES! Ostracod remembered the password to his TH-cam account!
Oh, and just saying, I think it'd be hella cool if Wheatsystem got ported to the DUO Adept.
Or duo mega
Maybe even the DUO Navigator (if you didn't notice, the website is run on a DUO computer)
This is awesome. Love the idea of a high-level environment to work on multiple architectures, reminds me of Java.
I was recently thinking of doing a similar idea (albeit MUCH less advanced), where I'd create an emulation layer running natively on a simple-as-possible Harvard 4-bit CPU with large program memory space (about 64KB to 128KB) to emulate a von Neumann 8-bit CPU architecture. The reason for a 4-bit architecture was to minimise on TTL chips as much as possible, with still the possibility of running 8-bit CPU programs. The emulation layer would chew up TONS of CPU time tho, I think the ratio of native 8-bit CPU to emulation 8-bit CPU instruction speeds was about 10:1 lol, making CPU-heavy programs extremely slow, but still theoretically doable.
Sounds just like what I have been designing!
welcome back to TH-cam! love your videos.
So that's what you where up to :^)
This is really cool dude. Can't wait to see what else you make!
This is a really cool project, I've been thinking of writing similar things. Did you explore any existing bytecode formats before you started on your own? I think WASM is a pretty good fit for such an OS, or maybe even some mini version of the JVM. The benefit of course being a massive pre-existing tool ecosystem (LLVM can produce WASM for example)
This is a cool project! I’ve been working on a 16 bit computer, kind of like the Amiga, using a 65c816, and I want to try porting WheatSystem to it, at least as an app running under it’s OS.
Inspired by your OS and having started growing cereal at home myself (rice to be exact), I might write a Rice OS.
BreadByteCode is sort of the same idea as "Machine Interface" of IBM's AS/400 computers. From small to really large all the AS400 machine run an emulator which runs the client programs and the OS. (as best I understand it)
Why do you always use throwaway batteries?
holy crap it's been 2 years since the last video!
Absolutely stunning!
I love ostracod computers
It's been a loong time!
How have you been?
computer,,...
this all sounds really cool! wheat- and breadsystem are v nice concepts and the demonstration was really interesting c:
I still have my Duo Decimal on my desk.
Amazing 🤣
But will it run on your relay computer? 🤣
Corn is best crop
Wheat is worst