f66 vs f77

[statespacedev]

wow - learning a lot over last few days:) will just copypasta some of the notes here to give a concise catchup

decwar is a fusion of fortran and assembly. roughly speaking, much of the real work happens in assembly, it's usually interacting with the high segment memory where the overall game state for multiple players/jobs lives. the fortran is somewhat of an i/o layer, interacting with the players. it's doing quite a bit of text i/o. now here's the thing. very roughly speaking the assembly code is 'six bit ascii' and expects the same with fortran, which means the f66 fortran66 ascii standard, hollerith characters, etc. f77 fortran77 ascii standard brought breaking changes with 'seven/eight bit ascii', character data type, suppression of hollerith, etc. here's the two relevant tapes for decwar.

tapes/BB-D480C-SB_FORTRAN10_V6.tap   f66  'six bit ascii'
tapes/BB-D480F-SB_FORTRAN10_V10.tap  f77  'seven/eight bit ascii'

we gotta care about that f66 tape. decwar is 'natively' f66. will discuss below 'porting' decwar to f77, it seems possible, but probly means breaking changes in the assembly code as well.

will be tracking all this with fuller notes here - https://github.com/drforbin/decwar/blob/minimalist/docs/sec8-f66-vs-f77.md

[statespacedev]

fwiw fyi - did indeed get decwar building and running on f10v10 f77 'ok-ish' - was actually playing with two players/jobs etc - but there's definitely i/o gremlins and looks like a major job to fix all that - to say the least:) set that aside in arc/utexas-f77 and focusing on f66 for now!

good question is which direction the 'math stuff' in docs/algebra and docs/inverse will go. f77 makes things 'easier' at least superficially... but does it make sense to 'stay old school' with f66, especially in light of 7090/7094, cdc, etc - in other words - for matrix inverse and simplex, fifties and sixties style - f66? or f77?:)

btw - highly recommended read!

[statespacedev]

can't resist - just read it, hahaha!!!