> I think we can include table like this one:
>
> https://web.scheme.org/
>
> But larger with more essential Scheme features, and those that implement
> more of them will be on top. More mature Scheme implementation will most
> likely implement all the features.
>
> For some people the features will not matter. Only the language of
> implementation and where it runs. For some they will pick the one that
> implement most of the features.
>
> We can do something like Wikipedia does for parser generators:
>
> https://en.wikipedia.org/wiki/Comparison_of_parser_generators
>
> we can indicate the supported feature with green and not supported with red
> and partial with yellow.

This is all reasonable.

But I think we should present some kind of summary by default so that
newcomers are not overwhelmed by too much detail.

It would be good to define some "profiles" of RnRS conformance. A
profile would be a group of features, and criterion for how fully each
feature is supported.

This will probably take a lot of experimentation to get right.