stringprep

Erik van der Poel erikv at google.com
Sun Dec 7 18:58:19 CET 2008


In Section 6 of Rationale...

http://tools.ietf.org/html/draft-ietf-idnabis-rationale-05#section-6

...the following text appears:

      For this case, the best practice will usually be to apply NFKC and
      case-mapping (or, perhaps better yet, Stringprep itself)

If the intent of the words "or, perhaps better yet" is to offer an
idea for better interoperability, may I suggest that Stringprep, by
itself, does not guarantee interoperability, because it offers
numerous choices, some of which are intended to be selected by a
"profile" (such as Nameprep)?

In addition, Stringprep refers to the old bidi rules, which have been
changed significantly in IDNA2008.

If we are to replace the text "or, perhaps better yet, Stringprep
itself" with some other wording, here are some alternatives:

"or, perhaps better yet, a new version of Nameprep, extended for
compatibility with IDNA2008"

"or, perhaps better yet, a profile of a new version of Stringprep,
extended for compatibility with IDNA2008"

"or, perhaps better yet, something akin to a profile of a new version
of Stringprep, one that takes IDNA2008 and new versions of Unicode
into account"

""

The last alternative is the empty string, i.e. remove the old text.

Erik


More information about the Idna-update mailing list