Liaison statement
Response to "Ecma TC39 Comments on RFC 4727bis" from 2013-11-25

State Posted
Posted Date 2013-12-10
From Group json
From Contact Paul Hoffman
To Group ECMA-TC39
To Contacts istvan@ecma-international.org
CcMatthew Miller
Pete Resnick
Barry Leiba
Purpose In response
Attachments (None)
Body
Thank you for your statement of concern about draft-ietf-json-rfc4627bis. The
chairs of the IETF's JSON Working Group were not aware of any serious concerns
that TC39 had with the update process. Over the past many months, we
repeatedly contacted ECMA and TC39 members about the JSON WG and
draft-ietf-json-rfc4627bis, but never received any significant reply. We know
that some TC39 members joined the JSON WG mailing list but did not voice any
process concerns in the WG -- or privately to the WG Chairs -- before or
during either of the Last Calls on the document.

In the future, it would probably be useful for Ecma and the IETF to have a
formal liaison relationship. We have heard that there were preliminary
discussions several months ago, but stopped short of a formal relationship
before the JSON Working Group was formed. Formalizing the liaison relationship
will help both SDOs communicate with each other and thus avoid late
surprises.

As to your specific requests:

- In IETF Last Call, there was consensus to change the definition of "JSON
text" in draft-ietf-json-rfc4627bis to match the one in ECMA-404. The latest
draft reflects that change. At this point, we believe that there are no
syntactic differences between the two specifications.

- The JSON WG decided to change the title of the document to better reflect
its content. The current document is much more than a media type registration:
it repeats the JSON syntax originally documented in RFC 4627, and has been
stable for many years, it is a discussion of the JSON semantics important to
freestanding encoders and parsers, it is a discussion of interoperability
issues that have been encountered since RFC 4627 and ECMA-262 5th Edition were
published, and it is a media type registration. Having a more accurate title
on the document will help readers understand its contents and the difference
between it and ECMA-404.

- After Ecma published ECMA-404, the JSON WG discussed whether to remove the
ABNF version of the JSON syntax that was established in RFC 4627 from
draft-ietf-json-rfc4627bis; it was decided not to do so. One reason is that
ECMA-404 uses "racetrack pictures" to define the syntax, whereas IETF
documents have traditionally used ABNF, and many developers have expressed a
strong preference for the ABNF. This might be considered simply a matter of
style, but it was deemed important by many WG members. We intend to keep the
discussion and reference to ECMA-404 in draft-ietf-json-rfc4627bis, even if
Ecma continues to choose not to reciprocate in ECMA-262 6th Edition, because
developers reading draft-ietf-json-rfc4627bis might indeed prefer the
racetrack pictures.

- A normative reference to ECMA-404 would be premature without a clear and
well-understood document management process. Historically, when someone
reading an RFC sees a normative reference to one version of another SDO's
standards, they tend to think the reference will apply to future versions of
that external standard as well. Given the closed process that resulted in
ECMA-404, it seems quite possible that Ecma could later make changes to
ECMA-404 that would have a negative effect on interoperability from the
Internet perspective. When the IETF normatively refers to other standards, it
tries to do so to standards that were developed with processes that are open
to discussion and contribution by anyone.

One of the positive products of a formal liaison relationship between the IETF
and Ecma would be better coordination of development of each other's
standards, which in turn could lead to more use of normative references in
future versions of documents.