Am Sa., 15. Aug. 2020 um 18:46 Uhr schrieb John Cowan <xxxxxx@ccil.org>:

> One of the reasons why there was so much resistance to R6RS from existing implementers was its rigidity, and I have reason to believe that the R6RS condition system was felt to be extremely bad in this respect.  R7RS has taken much from R6RS of great value, but condition creation is part of the core, and I have less than no desire to force implementers to change it.

Please see my question in this thread whether the R6RS condition
system can be implemented portably on top of what R7RS provides.

>> and a fast R7RS implementation built on top of the R6RS
>> implementation Chez like Racket on top of Chez would make perfect
>> sense.)
>
>
> +1.  I believe that such efforts are in progress.  You do want to replace psyntax with one that handles R6RS and R7RS equally, as in Larceny and some other systems, so it's not quite "on top of".

Racket uses Chez just as its backend. The frontend, including the
syntax expander, is Racket's one. Some other features like
continuation marks had to be wired into Chez as well.