RE: Namespace consistency

At 05:01 PM 9/11/98 PDT, Jim Whitehead wrote:

>The implication of the requirement in section 4.3 (no namespace gaps)
>combined with the requirement in section 4.1 that namespaces must behave as
>DAV collections even if they weren't created with WebDAV, is that yes, the
>resource identified by http://warlok.ds.boeing.com/~howie/ needs to be of
>type WebDAV collection if http://warlok.ds.boeing.com/~howie/a/b/c is
>WebDAV-compliant.

I agree that this is the implication of the protocol as it stands.

I would like to raise an objection to it, as it would prohibit one from
supporting DAV methods on anything smaller than the full namespace of a
server.    
It is often useful for various subtrees of a server's namespace to support
different sets of methods (sometimes via cgi scripts, without assistance
from the server maintainer) and there is no benefit to this restriction
(unlike the one requiring internal consistency).

I have been meaning to complain about this for a while but didn't because
it seems to be trivial.  I was having a hard time imagining anyone being so
picky as to insist on this point (and, perhaps, denying me the covetted
"100 % Pure DAV" sticker for my server?) but perhaps it's better to face it
now.  In fact, this very issue has been raised in some discussions internal
to Xerox.  People have said to me "We would like to support DAV methods in
this subtree but we don't want to support them on the root but the protocol
says we MUST do it, so can we still use DAV?".  So Howie isn't the only one
worrying about it.

I say, strike the restriction.

Received on Saturday, 12 September 1998 15:06:14 UTC