Stream: ideas

Topic: sister syntax

Ron Panduwana (Feb 22 2023 at 12:17):

Is there any plan for having an alternative syntax? e.g. Erlang has Elixir, Ocaml has Reason, even Java has Clojure etc. We can get more audience by providing alternative to Haskell-ish syntax. I suggest Sweet Expression which marries s-expression with something Python-like. Anyone interested?

Georges Boris (Feb 22 2023 at 12:30):

Apart from Reason, these languages are not really alternative syntaxes, right? They are complete languages that can interop with other languages through their runtime (just as Roc can do by the design of application <> platform).

Are you suggesting an alternative syntax for familiarity with C-family languages? Or something more dynamic, without the effect pattern and "free" FFI with the parent language? (like elixir with erlang and clojure with the java?) because then we're also missing other core parts of the language like purity.

Ron Panduwana (Feb 22 2023 at 12:44):

Purity definitely should be kept, so yes it's about familiarity gap/intuitiveness to non-Haskellers and in the case of sweet-expression, power (through macros)

Joshua Warner (Feb 22 2023 at 15:41):

Coming mostly from a "C-family" background myself, I think this could be interesting to experiment with.

Richard Feldman (Feb 22 2023 at 16:15):

Ron Panduwana said:

Is there any plan for having an alternative syntax?

the plan is not to do this :big_smile:

Richard Feldman (Feb 22 2023 at 16:16):

Elixir and Clojure are different languages from Erlang and Java, so I wouldn't say those are alternate syntaxes per se

Richard Feldman (Feb 22 2023 at 16:20):

historically I've seen a couple of things come out of alternate syntaxes:

• The alternate syntax is popular among some but not all people using the language, and it fractures the community - Reason is an example of this
• The alternate syntax sees very little use in practice - every project I can think of which fits the description "this does not change the semantics of the language but uses a syntax based around S-Expressions" falls into this category :sweat_smile:

Richard Feldman (Feb 22 2023 at 16:21):

so based on how such things have worked out historically, I think if we went out of our way to support this, we should expect negative results in practice

Richard Feldman (Feb 22 2023 at 16:21):

...which is why the plan is not to do that! :big_smile:

Joshua Warner (Feb 22 2023 at 17:59):

Yeah that's fair. To be clear, when I said "this could be interesting to experiment with" - I really meant insofar as this is easy+natural - and/or out-of-tree. As a concrete example, it could be cool to be able to write a roc editor plugin that lets me edit in a more familiar-looking syntax.

Last updated: Jun 16 2026 at 16:19 UTC