Partial answer https://github.com/bkw777/dl2/blob/master/co2ba.md
It only handles the simple case where there is nothing exotic or tricky about the .co, but is super easy and does handle all variants including nec. Unlike your co, it does not try to generate a single polyglot BASIC, you have to generate seperate nec and everything-else versions. -- bkw On Thu, Feb 12, 2026, 1:41 PM B 9 <[email protected]> wrote: > I created a single machine language .CO file > <https://github.com/hackerb9/crc16-modelt/blob/idtest/CRC16.CO> which > should run on any of the Model T variants (at least it does in emulation), > but now I’m trying to figure out how it would be loaded onto them. As a > binary it is not easy to transmit. > > Ideally, I’d have a BASIC loader that would POKE the program into memory, > but the problem with that is that my .CO file is already 1.5 kilobytes > long, so it barely fits into memory on an 8K machine. (Which, in itself > wonders me. Please see my calculation below and correct me if I’m > mistaken.) I believe a BASIC loader would expand the file size and simply > not fit. > > So my first question is: Is there something like TEENY for all Model > T’s,even for the less common variants like the Olivetti M10, that can > receive binary files over the serial port? > > My second question is: Am I confused about the 1.5K limit on > cross-platform .CO files? I’m fairly new to assembly language on these > devices and have made some guesses based on what I’ve been observing in the > Virtual-T memory viewer. Let me state my presumptions explicitly, so you > all can help me figure out where, or if, I’m going wrong. > > 1. > > A single .CO program can run on the Kyocera Kyotronic 85 and its > sister computers: TRS-80 Model 100, Tandy 200, Tandy 102, Olivetti M10, NEC > PC-8201A, and NEC PC-8300. > 2. > > All the Kyotronic kin must store a .CO file *twice* in memory: Once in > the filesystem and a second copy in high memory when it is loaded through > LOADM, RUNM, or by selecting it from the MENU. [Side question: Does > distribution on cassette get around this limitation? Do all machines have > the same audio encoding format?] > 3. > > The area in high memory which the .CO file is copied to before it runs > must be reserved using the command CLEAR 256, *x*, where 256 is the > space reserved for strings, *x* is the lowest memory address the .CO > file can occupy, and the builtin variable *MAXRAM*-1 is the highest. > 4. > > I’m defining *MINRAM* as the lowest address which can be reserved > using CLEAR. The value of MINRAM is higher on machines with less memory. > 5. > > To run cross-platform, we’d need to know the largest MINRAM and > smallest MAXRAM across all the Kyocera kin. The Tandy 200 has a MAXRAM of > 61104. The 8K Model 100 has a MINRAM of 58044 (with no files and setting > string space to 0). > 6. > > Thus, the largest cross-platform machine language file would be: > > [image: \frac{(61104-58044)}{2} = 1530 \text{bytes}] > > and the space from 59574 to 61103 should be reserved for the .CO file > using the command > > CLEAR 0, 59574 > > > What is the right way to load a larger file? Or is it not possible and I > need to solve my problem in a different way? (I’ve already had to compress > the data strings to get it to fit at all. I could do more, but the code > would be uglier and harder to extend in the future.) > > Thanks all for any help or suggestions, > > —b9 >
