Re: Multiple Content-Location headers
Einar Stefferud (Stef@nma.com)
Tue, 13 Jan 1998 15:26:31 -0800
Hi Jim -- Speaking as MHTML Chair, I must observe that you sound a lot
like I sounded when I first looked at a lot of the stuff going on in
HTTP;-)... Just ask Larry Masinter;-)... He bore the brunt;-)...
Since then MHTML and HTTP have been cooperating at least at the Chair
& Editor levels. I expect that more of the MHTML participants have
been aware of this inter-WG negotiation because most of it was
broadcast to the MHTML list. Larry Masinter and Roy Fielding were the
primary negotiators on the HTTP side. I expect that most of the
issues seemed to be minor adjustments not needing broad HTTP WG
efforts, but they were more important to MHTML at the time.
So, welcome to the club;-)... Some of us have been waiting a long
time to get this MHTML/HTTP discussion under way;-)... We really do
need to sort this out among us and try to find Rough Consensus across
both WGs, if not more WGs.
Now then, a detail:
How are you going to handle HTTP retrieval of archived ever-changing
weather maps (that normally change within a given URI which stands for
"the current weather") that have been archived as compound MHTML "MIME
Objects"? If you don't retrieve each of them with all of their
archived parts, how are you ever going to reconstruct what the weather
was at any time in the past?
So, having found at least one case that is clearly going to make good
use of MHTML for archiving and transfering compound web objects, how
are we going to enable this across the whole spectrum of transports
(HTTP, RFC822, FTP, SneakerNet, CDROM, EtcNet) for such compound WEB
objects whose parts need to be bound to each other?
Best...\Stef