Re: News and nntp URIs

Charles Lindsey wrote:
  [Tom Petch wrote]
>> s8.3 I cannot quite follow, a bit more background might be useful

It's an erroneous IANA registration.  Henry Spencer, the author of 
"son-of-1036", introduced MIME into NetNews, and while he was at it
he registered message/news.  It's an 8bit variant of message/rfc822,
the USEFOR WG will formally deprecate it in I-D.ietf-usefor-usepro:

He also registered news-message-id as message/external access type,
but somehow that ended up in the wrong registry for *media* types
<> instead
of <>.

Fixing a very old registry error, it's not important *where* that's
done, but it has to be an RFC (required by RFC 4289).  Henry Spencer
registered news-message-id before the RFC 2048 rules (now RFC 4289)
existed.   Don't use application/news-message-id, it's as bogey. :-)

>> I see half a dozen places where the English might be smoothed a
>> little.

Charles already proposed various improvements, whatever is left is
clearly my fault, please tell me what's still on the wrong side of

> This draft has been discussed on the ietf-nntp list

Yes, I answer Clive's articles later, I waited for the publication
of RFC 5064 for the next round.  And I'll answer here, we obviously
all read this list.  I'm quite willing to remove wildmat completely,
the percent-encode-? effect is admittedly odd.  But I can't remove
"*" wildcards, they already existed (as note) in the Gilman draft,
and they are implemented.  Besides Russ and somebody else wanted a
general "*" wildcard beyond the single use case in RFC 1738.

Your 2822upd objection was a waste of energy, of course I'll update
the URI draft IFF (and when) RFC.ietf-usefor-usefor is updated, it
is *the* normative reference.  Apart from RFC 3986, showing how
STD 66 and RFC.ietf-usefor-usefor work together is the point of
the URI draft, based on common practice outlined in Gilman's I-D.

Talking about wildmats is certainly not the point, I can drop it.


Received on Saturday, 29 December 2007 00:04:36 UTC