As I didn't get any responses, in particular, no negative ones, I have
implemented the proposed refinements.  Please check the newest draft.

For a follow-up draft, I am thinking of improving the R6RS syntactical
layer to bring it on par with the procedural layer, allowing it to
define more than one constructor for the same record type.

Am So., 23. Okt. 2022 um 12:04 Uhr schrieb Marc Nieper-Wißkirchen
<xxxxxx@nieper-wisskirchen.de>:
>
> I want to propose the following amendments to the R6RS record facility
> as used in this SRFI to address the concerns voiced in SRFI 99. This
> follows the reconciliation wish prominently embedded in this SRFI's
> title.
>
> - The define-record-type syntax binds the <record name> to a
> (syntactic) keyword.
> - A record constructor descriptor, which holds an associated
> record-type descriptor, can be used in place of a record-type
> descriptor, where it stands for the associated record-type descriptor.
> - As an (identifier) expression, the record name keyword evaluates to
> the same record constructor descriptor, to which the expression
> (record-constructor-descriptor <record name>) evaluates.
> - The parent clause in a define-record-type form can also be used with
> an expression evaluating to a record constructor descriptor instead of
> just a record name.
> - The language "record descriptor" is used for "record constructor descriptor".
>
> I am looking forward to hearing your comments.