Re: RFC2822bis, Message-IDs and USEFOR

["Frank Ellermann" <nobody@xxxxxxxxxxxxxxxxx>]

Harald Tveit Alvestrand wrote:
 
> Unless somehow a horde of people with lots of energy and competence 
> appears on the mailing list, I'm not going to entertain proposals for 
> reopening USEFOR.

FWIW, here's what I wrote yesterday to Lisa about this issue:

| Meanwhile I realized that replacing the normative RFC.usefor-usefor
| reference by 2822upd cannot work without reinventing the (USEFOR)
| wheel to some degree - posted on the general list yesterday.
[*]
| For 2821bis and other drafts it is difficult to decide.  But for
| RFC.usefor-usefor and the news URIs any "last minute" attempts of
| "2822upd upgrades" should be under your control.
|
| A simple recipe could be "WONTDO", 2822upd comes too late, and it
| is too easy to get "last minute" ABNF changes wrong.

*: <http://article.gmane.org/gmane.ietf.general/29592>

> Just finding the energy to call the question on the last
> outstanding USEPRO question is proving too much for these chairs.

Some foreseeable 2822upd simplifications are IMO more relevant for
EAI and mailto-bis than for USEFOR.  Mailto-bis is in the critical
path for mailto-eai + eai-mailing-lists.

Even if hordes of folks with lots of energy show up - at some point
in time somebody needs to figure out who got <toplabel> right, 4408
or USEFOR.  If ICANN seriously wants weird one letter A-<toplabel>s 
USEFOR got this "wrong" :-(  

Neither 4408 nor USEFOR mention subtle details about "xn--"... vs. 
2ALPHA "--"... labels, let alone "0x" HEXDIG.  I'm not aware of any
RFC mentioning 0x0... up to 0xf... as "bad" labels.

If RFC.usefor-usefor goes to DS these details can be fixed, that 
obviously won't happen before ID.usefor-usepro is an RFC with two
interoperable implementations.

 Frank