Early Review of draft-dnoveck-nfsv4-internationalization-01
review-dnoveck-nfsv4-internationalization-01-i18ndir-early-williams-2020-06-01-00

Reviewer: Please consult with Pete Resnick on how to do the review.

Background info from the author:

This document is part of an effort to produce an rfc5661bis.  When looking at the work to be done we found that certain areas (e.g. internationalization, security) would be better dealt with on an NFSv4-wide basis. In the case of internationalization, this was because the treatment in rfc5661 was based on that in rfc3530, which was based on stringprep.  That treatment, apparently,it was dictated by the IESG as part of approving rfc3530 and then ignored by implementers.  As a result, we had to completely rework internationalization for NFSv4.0 in RFC7530, while NFSv4.1 remained as it had with an internationalization section implementers rarely looked at and never implemented.

Although this is explained in the draft, I think I need to mention here that the constraints on NFSv4 implementations (in terms of external APIs and compatibility with/acces to existing file systems means that the IETF and the working group cannot direct the handling of UTF8,  particlarly with regard to normalization, normally expected.

To get back to rfc5661bis, we decided to write an NFSv4-wide draft adopting the treatment of RFC7530 for all NFSv4 minor versions.   Then we ran into the problem that RFC7530 had  been written assuming use of IDNA2003 so that the text of RFC7530 ran into idnits problem since it referred to obsoleted IDNA RFCs.   As a result the transition to IDNA2008 is one of the major issues that need to be reviewed in draft-dnoveck-nfsv4-internationalization-01.