New notation for records?

[AntC2]

@atravers

what everyone really wants is C-lassic record notation

Not everyone: I do not want nested records. (And there’s nothing secretly-in-my-heart I’m failing to admit to myself.) I’ll illustrate:

data Country = Country {name1 :: String, company :: Company} deriving Show
data Company = Company {name2 :: String, boss :: Employee, car :: Car} deriving Show
data Employee = Employee {name3 :: String, age :: Integer, empCar :: Car} deriving Show
data Car = Car {name4 :: String} deriving Show

This illustrates the classic/naieve hierarchical data structure that Commercial IT abandoned in the '80’s. Why does Haskell persevere with that thinking?

f :: Country -> Integer
f = age . boss . company

Wrong. Join Country to Company to Employee. You now have a flat/wide tuple/record [**] grab the age from that. Or update the age into it.

[**] in which naming the same thing with the same name will make joining much clearer, no clutter of different names for the same thing; no need to RENAME or SELECT-by-name one of the same things.

Now admittedly update-through-Join can be hazardous. (But not in straightforward cases like this example.) This is a reasonable area for research, you’da thought Haskell would be right at the forefront. But no. GHC is not even at first base,

Addit: IOW what @BurningWitness said. That post arrived as I was typing.