[Top] [All Lists]

Re: informal last call for draft-duerst-archived-at-06

2006-11-08 15:11:37

Frank Ellermann wrote:
Graham Klyne wrote:

I think a distinction needs to be drawn here between *registration* and

Point - that solves my first nit, X-Archived-At can be registered, but not
with "Status: standard".  Martin's draft proposes a "Status: deprecated",
so far that should be okay.

I see no reason why the X-Archived-At header should not be included in
the provisional registry -- indeed to do so seems to me to be entirely
in accord with that registry's purpose.

ACK, your concept is better than my "impossible".  With a provisional entry
and "Status: deprecated" at the same time folks are free to (ab)use this
X-header-field for completely different purposes.  If they're smart they
don't do this or at least wait some years, but that's another issue... :-)

:)  No argument from me.

=== 2nd point ===
the planned EAI experiment might still limit its scope to local parts
in addresses (excluding Message-IDs).

Archived-At has "intended status: standards track", not "experimental".

Responding to the last paragraph only -- I'm not sure how it relates to
the preceding text -- I think that the "experimentation" has been fully
conducted by W3C and their mailing lists using X-Archived-At, so I'm not
sure why one would consider "experimental" status at this time.

The intended status of the I-D is PS, "Status: standard" for Archived-At
is perfectly okay (modulo last call + approval).  But setion 2.4 in the
I-D obviously talks about EAI, and unlike Archived-At EAI is (1) far from
ready for prime time, (2) experimental (when ready).

Therefore I think that section 2.4 is premature, and in essence it simply
says "as soon as IRIs are possible message headers producers can do what
RFC 3987 says", nothing special so far.

But at this point in the game it's IMO too early to talk about IRIs in
message headers (Archived-At is for mail + news, not http etc.), it would
only cause unnecessary confusion.  Software producing IRIs can as well
produce URIs, the latter works with more clients.  Clients are always the
most backwards part, where it takes a decade until they learn new tricks.

Clicking the "about netscape" button in my browser:  1994-1997.  A decade
is already optimistic... :-)  Let's start with Archived-At IRIs when EAI
is on standards track (maybe late 2008).

Ah, I see what you mean.  I hadn't appreciated you were talking about section
2.4, which I hadn't fully absorbed.  I don't have any cogent view on this, so I
should refrain from further comment.  (Except to suggest that maybe the section
could be marked as non-normative?)


Graham Klyne
For email:

<Prev in Thread] Current Thread [Next in Thread>