Re: MHTML/HTTP 1.1 Conflicts

>If the HTTP group has decided to remove Content-Base, are the
>same reasons valid for removing Content-Base from MHTML? The
>difference is that when you send by e-mail, you have no request
>URI, so this might be a reason why Content-Base is more important
>in e-mail than in HTTP.

Yes, the same reasons are valid for MHTML.  In fact, they came out
of our earlier discussion regarding the confusion between Content-Location
and Content-Base within a hierarchical context.

The reasoning is that the only principal who is capable of knowing
whether or not the embedded relative URLs of a document are relative
to a base URL that is different from the document location (the only
reason we have Content-Base) is the person/process that created that
document.  Since it is reasonable to assume that any media type
capable of using relative URLs in that manner will also have its
own means of assigning a base URL (e.g., the BASE element in HTML),
we don't need a special field for that purpose at higher levels in
the document hierarchy.  Furthermore, since an embedded BASE is known
to be more reliable for this purpose than any other mechanism, we
don't need Content-Base for the single-document case either.  In addition,
it greatly simplifies the problem of "how to get the Base URL", since
we remove any ambiguity between Content-Base and Content-Location
precedence.

Since it is unlikely that anyone is currently sending MHTML documents
with Content-Base (they only need Content-Location), I think
that MHTML can safely drop Content-Base at this point.

....Roy

Received on Wednesday, 28 January 1998 14:25:55 UTC