…tomics

it’s quite important that
* there not be empty cells in the head or body between contentful cells,
* there not be `""` below the `DECIDE` --- if there are `""` below the `DECIDE`, then the stuff below will get parsed as distinct Hornlikes but without the givens, and the only way to then figure out what the original givens were will be very fragile

…e: Check upfront for wehther there are non-text mtexpr variable names in the GIVENs; raise a `dispute` if so and print warning as comment in resulting .le

…ics` and `optics`

…eamline the code (remove overly complicated typeclass typeapplications stuff) while still getting compiler support for subtypes of RPRel, by using GADTs and my own RPRel-esque type

2. Add support for arith comparison

…not working on it right now

…nt places in NLAs (instead of just 'a var') and for more nuanced filtering downstream

…tification

…have to escape metachars in Pred)

…tions over concrete data structures

… NLAs

…ashing NLAs with the NLATxts and not the bases)

…hether rly need to escape stuff, and what to do when there is both an inferred NLA from is-num as well as another NLA with the same kind of pattern inferred from another use

…g NLA has more vars than the subsumed one), among the non-lib NLAs. (Will need to handle subsumed but same-num-of-vars NLAs separately.)

…port the `NLA` record without exporting the accessors

…LENLA`; and add `regextravifyNLASection :: T.Text -> Maybe [RegexTrav]`

…eriod, just in case

…subsumed (etc) NLAs into one place (in LE.hs), to make code more modular and readable

…t-up-to-varnames (with same number of vars) NLAs

… var names (and have the same number of vars) as comments in case the user didn't mean for them to be removed

… of having `isReg` and `isHlike` functions, to avoid confusing people