You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We want bindings/variables, obviously (see #19). Another is the ability to define new procedures. This will require a procedure value. I am thinking that a new procedure could be defined like...
(define times-two (function (n) 2*n))
(Related to #19, I prefer define as a keyword, which is what I'm used to in Racket. It would be good to set as a separate keyword, so there is a separation between declaration/initialization and changing of values, like := v. = in Go)
Furthermore, all procedures in Limp should be defined as procedures in the environment, including (especially) built-in ones. But, there should be room for some distinction between procedures defined purely in Limp and ones that rely on "native" (Rust) code. Because obviously there has to be some base of things defined in Rust, just like there is for say, Ruby. For this, some kind of native interface would be useful. And there needs to be a way to define native-based procedures as individual modules, instead of putting them all in one giant switch statement.
Needless to say, figuring out how to implement all of this is rather complicated, and could be subject to a future hackathon/project. Some research is required.
We want bindings/variables, obviously (see #19). Another is the ability to define new procedures. This will require a procedure value. I am thinking that a new procedure could be defined like...
(Related to #19, I prefer
define
as a keyword, which is what I'm used to in Racket. It would be good toset
as a separate keyword, so there is a separation between declaration/initialization and changing of values, like:=
v.=
in Go)Furthermore, all procedures in Limp should be defined as procedures in the environment, including (especially) built-in ones. But, there should be room for some distinction between procedures defined purely in Limp and ones that rely on "native" (Rust) code. Because obviously there has to be some base of things defined in Rust, just like there is for say, Ruby. For this, some kind of native interface would be useful. And there needs to be a way to define native-based procedures as individual modules, instead of putting them all in one giant switch statement.
Needless to say, figuring out how to implement all of this is rather complicated, and could be subject to a future hackathon/project. Some research is required.
(Other potential future features: strings, arrays, maps.)
The text was updated successfully, but these errors were encountered: