Big thanks to Arthur for the reacting quickly and suggesting the
last-call week. The 90-day deadline for this SRFI is now up and I
don't have anything to add or change. Nobody else has commented since
April either. This is not surprising as the syntax was thoroughly
discussed with many experienced Schemers on srfi-discuss prior to
submitting the SRFI.

This draft does not introduce any material changes or additions to the
specification or rationale from the first draft. However, both have
been clarified extensively (mainly by pointing out more edge cases). A
sample implementation and acknowledgements section have been added.

The diff can be quite difficult to read due to the extensive editing
and especially because the new draft was re-formatted using HTML Tidy
right before submission. I recommend simply reading the latest version
at <https://srfi.schemers.org/srfi-169/srfi-169.html> with help from
the following guide.

Section-by-section changes between drafts #1 and #2:

- Abstract                               (unchanged)
- Rationale
   - How many digits per group            (unchanged)
   - What separator character to use      (unchanged)
   * Potential ambiguity between numbers  (many clarifications)
   * Printing numbers with underscores    (completely new section)
* Specification                          (many clarifications)
- Examples                               (unchanged)
* Implementation                         (completely new section)
* Acknowledgements                       (completely new section)

Thank you to everyone who has helped with this SRFI :) In case you
have any comments, they are still welcome.

The sample implementation passes its test suite on Chibi, Gauche and
Kawa. Git: <https://github.com/lassik/srfi-169-sample-implementation>