Thank you for your thoughts, Shiro. I haven't replied yet because I
haven't found a better name than "parameter-like object". In fact,
some Schemes call what SRFI 226 calls a parameter-like object a
parameter, so using something completely different may let SRFI 226's
terminology diverge too much from existing terminology.

What would you think of "procedural parameters" instead of "parameter-like"?

A "procedural parameter" would not be a parameter, but, at least in
mathematics, this is not uncommon. A "manifold with boundary" is not a
manifold (which is a manifold with boundary with empty boundary...).

Am Di., 27. Dez. 2022 um 20:13 Uhr schrieb Shiro Kawai <xxxxxx@gmail.com>:
>
> I'm updating Gauche documents reflecting recent changes to support srfi-226, and wondering if there's a better name for parameter-like objects.  "Parameter-like" does capture its behavioral aspect, but explaining it as "it's not a parameter, but somewhat behaves like a parameter, though behaves differently in corner cases" seems not straightforward.  Explaining something (X) by saying "X is not Y but like Y" doesn't seem good.
>
> Instead, we may frame it as a protocol to manage dynamic state.  They can not only be a value-holding passive object, but also do some active work (e.g. external resource acquisition and release) when dynamic state is changed.  And then we explain that it superficially resembles a parameter but needs to be managed by 'with', and may behave differently when composable continuation is used.
>
> I haven't come up with a nice name, though.  "Dynamic state" would be too broad.  "Dynamic state handler" may be too long.
>
> Any opinions?
>
> --shiro
>