On 6/29/21, 2:14 PM, "Aaron Hill" <[email protected]> wrote:
On 2021-06-29 12:45 pm, Carl Sorensen wrote:
> I believe that the logical implications of side-effects are not very
> concerning when local variables are set!, so there's probably very
> little or no downside to the set!s in Aaron's code. In fact, it's
> possible that Aaron's code will use less heap space by re-using class
> instead of adding a new variable grob-class-name. But I can't help
> myself; I want to avoid the grief Sussman would give me over two
> unneeded set! calls.....
I prefer that pattern of set! calls because it helps you avoid nesting
S-expressions too deeply, which I feel greatly improves readability and
maintainability. It is the same reason I bothered to define the
add-class-name procedure, so that it would not be nested within the
engraver definition itself, better separating concerns.
Additionally, I have taken flack for pushing some of the more functional
elements of Scheme programming. (Mention the word "lambda" once, and
everyone assumes you are a pure functional zealot.) Regardless, I have
tended to favor a more imperative approach lately, something that is
likely to be more consumable for users of LilyPond.
I don't disagree with your assessment.
Thanks,
Carl
