If an opinion matters, I am strongly against anything non-ascii anywhere
in the code, except, maybe, static strings.

Saying that, however, I do use Emacs fontification to substitute fancy
symbols in place of lambda, define, set!, set-car, set-cdr, and some
other built-ins.

In my experience it _does_ improve readability, it's just that source
file encoding is not the place to define it.

Maybe you would like to think of something like CSS for Scheme?
Something that does not influence the way the code is interpreted, but
influences the way it looks like?

You could supply a "canonical" "stylesheet", but each programmer would
be free to override it.

Lassi Kortela <xxxxxx@lassi.io> writes:

>> (BTW, if "lambda" is too long to write---I'm using ^ in place of lambda and it's
>> quite handy.  Of course, these days, λ is another plausible choice.
>
> Is Scheme portability at a good enough level that we could now do a trivial SRFI for
> Unicode lambda?
>
> (define-library (srfi nnn)
>   (export λ)
>   (import (scheme base))
>   (begin (define-syntax λ (syntax-rules () ((λ . rest) (lambda . rest))))))
>
> The nice thing about the lambda character is, it's one Unicode codepoint, so it's
> oblivious to which normalization form is being used. Using Python 3:
>
>>>> unicodedata.normalize("NFC", "λ").encode("UTF-8")
> b'\xce\xbb'
>>>> unicodedata.normalize("NFD", "λ").encode("UTF-8")
> b'\xce\xbb'
>>>> unicodedata.normalize("NFKC", "λ").encode("UTF-8")
> b'\xce\xbb'
>>>> unicodedata.normalize("NFKD", "λ").encode("UTF-8")
> b'\xce\xbb'

--
Your sincerely,
Vladimir Nikishkin (MiEr, lockywolf)
(Laptop)