Perhaps it is something like RADIX 50 or RADIX 100?
Jeff Birt From: M100 <[email protected]> On Behalf Of Joshua O'Keefe Sent: Wednesday, July 31, 2024 5:55 PM To: [email protected] Subject: Re: [M100] Double precision format (N82 BASIC) and inaccuracy from simply saving a file On Jul 31, 2024, at 3:22 PM, B 9 <[email protected] <mailto:[email protected]> > wrote: Now, I have some questions: Is this well known? Is it documented anywhere? Was this implementation something Microsoft did or was the floating point math one of the additions NEC made for N82 BASIC? Is it possible for there to be a loop such that the number would never stop changing? Hi b9, I've got a bit too much of a raging headache (and diminished 8080 skill) to be entirely sure this is *exactly* it, or what the heck it's trying to do, but I think the original implementation itself is somewhere here: <https://github.com/z88dk/techdocs/blob/6a97f97b9f87fbadf08ab3c1e8d29bd6be28e9d9/targets/m100/pc8201.asm#L1584> I know zx70's commentary on the 8201 disassembly isn't as stellar as the 100/102 disassembly, but it's still at least somewhat readable. I hope this helps you along your quest for your (wonderful, thank you!) tokenizer's advancement. I've seen in general you've been shooting for a true-to-memory—or at least true-to-freshly-loaded, barring line relocation—design, so following the original even if it's irrational and nonlinear and absurd could be about all you might expect to do.
