My idea was that since Gambit (and other systems) can have its own implementation of the SRFI, this Gambit-specific implementation
can support current Gambit features of 'pp' as implementation-specific extensions without breaking compatibility. If the
specification says something that makes it impossible, I am willing to revise it, because that was its original intent.

On the other hand, I can make 'pprint' the base function, getting rid of 'pp' name altogether. This, of course, may cause
incompatibility with some other implementations, but my bigger concern is that I would need to rename the procedure that follows
R7RS 'write' rules w.r.t. circular structures: a procedure that just follows whatever the (dynamic) parameters tell it, and the
procedure that guarantees printing circular structures without hanging should be named differently.

-Sergei

-----Original Message-----
From: Marc Feeley
Sent: Friday, May 15, 2026 9:55 AM
To: Sergei Egorov
Cc: xxxxxx@srfi.schemers.org
Subject: Re: Symbol "pp" is already used by Bigloo, Gambit, MIT and STklos

> On May 14, 2026, at 4:19 PM, Sergei Egorov <xxxxxx@acm.org> wrote:
>
> It is more or less customary for SRFIs to propose names already in use for features that lack standardization. In this case, the
> proposed definition for 'pp' can be easily made backward-compatible with the current definitions. Implementors can do the
> following without breaking their existing code base:
>
> Bigloo: (pp obj [output-port])   -- can be just extended in a backwards-compatible way
> Gambit: (pp obj [output-port [readtable]])  -- can be extended to allow an optional readtable between port and keyword arguments

Gambit’s pp and pretty-print procedures only support an optional output-port second parameter. The readtable, which has an effect on
the printed representation, is attached to the port. For example:

> (output-port-readtable-set! (current-output-port)
                               (readtable-max-write-length-set (current-readtable) 20))
> (pretty-print (list (iota 10) (iota 40) (iota 100)))
((0 1 2 3 4 5 6 7 8 9)
  (0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 ...)
  (0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 ...))

Moreover, the pp procedure is meant for debugging, where the short name is convenient. It behaves slightly differently from
pretty-print: when the parameter is a procedure, it pretty-prints the source code of that procedure, if available:

> (define (foo x) (+ x 1))
> (pretty-print foo)
#<procedure #2 foo>
> (pp foo)
(lambda (x) (+ x 1))

Please don’t use pp as the name of the main pretty-printing procedure. This would complicate debugging when this SRFI is imported.

Marc