> On Feb 15, 2021, at 5:22 PM, Lassi Kortela <xxxxxx@lassi.io> wrote:
>
> On a related note that has less to do with the essence of Scheme but a lot to do with the essence of RnRS: there are some additional features that one would expect from implementations of the recent RnRS reports.
>
> The most significant change from R5RS to R7RS was the addition of a library system. IIRC there have been implementations billed as R7RS with little to no support for `define-library`. The main point of using R7RS over R5RS is to be able to share portable code over the internet as libraries, so lack of this feature is missing the main attraction of the report (and with substantial ramifications for practical programming).
>
> Likewise, R6RS without `(library ...)`, syntax-case macros, and the standard condition hierarchy would probably not have much to do with the spirit of R6RS.
>

I totally agree and this is why assigning a meaning to “Scheme”, “R5RS”, etc is important.  If labeling is not done in a disciplined way the labels don’t mean anything to those who have a need to know the level of conformance for portability reasons.  And that should be everyone interested in writing libraries of Scheme code to share with others.

Marc