As few peo­ple have asked for my opin­ion. Here it is in very con­cise form.

It de­pends on your goal and it is all very de­bat­able.

To fun­da­men­tally un­der­stand pro­gram­ming, I like a ‘barbell’ strat­egy. One lan­guage close to the ma­chine and one that is very ab­stract:

• C (the ma­chine)
• Haskell (the ab­strac­tions)

Bonus shouts to Racket (macros), Prolog (logical pro­gram­ming) and Erlang (concurrency).

The other lan­guages are prag­matic com­pro­mises of the fun­da­men­tal con­cepts that are learned more purely with the lan­guages above this line.

F# is a good com­pro­mise be­tween ma­chine and ab­strac­tions and might be a good way to kill two birds with one stone.

To be pro­duc­tive im­me­di­ately and find (easily?) re­mote (low?) pay­ing jobs as web de­vel­oper:

• HTML (not re­ally a lan­guage)
• CSS (not re­ally a lan­guage)
• Javascript

Bonus shout to Typescript.

To work in the cor­po­rate world for an ob­scene amount of money:

• C#
• Java

Bonus shouts to Powershell and Bash for au­toma­tion; Rust to look cool to the hir­ing man­ager.

To work as a data sci­en­tist (AI AI AI …):

• Python
• C++

… And a good math back­ground. Bonus shouts to R and Julia.

There are fan­tas­tic books on all these lan­guages. Everyone has his fa­vorite.

I also have no idea what is best to teach young peo­ple as I have never done it.

They prob­a­bly pre­fer to build funny stuff (i.e., games) which is not triv­ial with my ‘fundamental choic­es’ of C and Haskell. Compromises need to be made.

I know there is a good Racket book along those lines. Perhaps there are oth­ers. I have not in­ves­ti­gated the mat­ter.