Recursive look up of base in outer headers

Jacob Palme <> Sat, 30 August 1997 09:33 UTC

Received: from cnri by id aa02490; 30 Aug 97 5:33 EDT
Received: from (services.Bunyip.Com []) by (8.8.5/8.7.3) with ESMTPid FAA19508; Sat, 30 Aug 1997 05:36:25 -0400 (EDT)
Received: (from daemon@localhost) by (8.8.5/8.8.5) id FAA14339 for uri-out; Sat, 30 Aug 1997 05:12:50 -0400 (EDT)
Received: from (mocha.Bunyip.Com []) by (8.8.5/8.8.5) with ESMTP id FAA14334 for <>; Sat, 30 Aug 1997 05:12:48 -0400 (EDT)
Received: (from daemon@localhost) by (8.8.5/8.8.5) id FAA10179 for uri@services; Sat, 30 Aug 1997 05:11:23 -0400 (EDT)
Received: from ( []) by (8.8.5/8.8.5) with ESMTP id FAA10176 for <>; Sat, 30 Aug 1997 05:11:20 -0400 (EDT)
Received: from [] ( []) by (8.7.1/8.7.1) with ESMTP id LAA24117; Sat, 30 Aug 1997 11:12:36 +0200 (MET DST)
X-Sender: (Unverified)
Message-Id: <v03007800b02cd020a684@[]>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Fri, 29 Aug 1997 21:34:58 +0200
From: Jacob Palme <>
Subject: Recursive look up of base in outer headers
Cc: IETF working group on HTML in e-mail <>
Precedence: bulk

This message is sent to the uri mailing list, since this is the list
in which the work of updating RFC 1808 is discussed (IETF draft
"draft-fielding-url-syntax"). I am cc-ing this message to MHTML,
the mailing list for the IETF working group on sending HTML in e-mail.

The reason for this message is that we have noted a contradiction
between RFC 1808 and our proposed standard (RFC 2110: MIME E-mail
Encapsulation of Aggregate Documents, such as HTML (MHTML)).

The contradiction is thar RFC 1808 as well as draft-fielding-url-syntax
say that if there is no base in the HTML text, and no base in its
content heading, then you should recursively look at surrounding
content headings to find a base. RFC 2110, on the other hand,
says that you should only look for a base in the immediatly
surrounding header, not go out to surronding multi-part headings.

The MHTML group, at is meeting during the IETF meeting in Munich
a few weeks ago, decided that it preferred RFC 2110 to stay as
is written and not allow recursive lookup of bases. However, we
were during that meeting not aware that an already published
RFC, 1808, says otherwise.

The reason the MHTML group did not want this was mainly the
complexity, for example, if there is a Content-Location on an
inner heading, and a Content-Base on an outer heading, which
has priority?

I personally have become convinced that it is better to do as
RFC 1808 says, allow recursive lookups. The reason for this is
that when you really need an absolute URL, you should be allowed
to use all reasonable options to find one. If this is said, we
must clarify the ambiguity by saying explicitly that bases
in closer Content-Location headers take precedence over bases
in outer Content-Base headers.

If you reply to this message, cross-post to,
since otherwise I will not get your reply.

Jacob Palme <> (Stockholm University and KTH)
for more info see URL: