Am Mi., 17. Nov. 2021 um 09:52 Uhr schrieb Marc Nieper-Wißkirchen
<xxxxxx@gmail.com>:

> (2) The entry for the `dtd` syntax should be specified more formally. `procname` should be an `<expression>`, which is then evaluated to a procedure.  Furthermore, I suggested a different macro that is now specified in the SRFI.  The version in the SRFI is just a trivial wrapper around `make-dtd` (which makes it questionable why it is needed in the first place) and so sentences like "The macro may also verify that the proc-ids are valid, that there are no duplicates, etc." do not make much sense because the same job can be done by `make-dtd` with the same efficiency.
>
> What I did suggest was a macro
>
> (dtd (<kwd> <expression>) ...)
>
> where each <kwd> is an identifier (!) whose syntactic binding (!) is the same as the syntactic binding of `dictionary?(-id)?`, `dict-alter(-id)?`, etc.  This way, the checking can be done early at expansion time.

This hasn't been fixed yet, I think.