- From: Yaron Goland <yarong@microsoft.com>
- Date: Fri, 30 Jan 1998 16:06:57 -0800
- To: "'WEBDAV WG'" <w3c-dist-auth@w3.org>
Up until now we have used various URIs to name DAV defined properties and XML elements. To date we have used http://www.ietf.org/standards/dav/xxx and are currently using http://www.iana.org/standards/dav/xxx. Both names were arbitrary and were chosen based on the assumption that it would eventually be possible to register URLs with either the IETF or IANA. In fact such a project has been started by IANA but it is in its very early stages and is not yet ready to handle registering XML schemas or DAV property names. Furthermore there are very serious issues regarding who owns the schemas, who has the right to change them, what they should be, what information they should provide, etc. that are very far from being resolved. As such I propose the creation of the new DAV URI namespace. This namespace is NOT resolvable to anything. It can ONLY be extended by having the new names published as an RFC. Its only purpose is to create a unique namespace for use by the DAV standards. I imagine three questions immediately come to mind: 1) Wait, what if someday I do want to be able to resolve the DAV names into a retrievable schema? No problem. The way we name everything in DAV is through the XML namespace extension. The form of that extension is <namespace name=DAV:xxxx href=http://www.foo.bar/blah as="dav">. The name attribute is required and is the unique name of the namespace, it MAY be a resolvable URI but does not have to be. The href attribute is completely optional but if used it MUST point to a resolvable URI which can provide the schema. Thus when the IANA schema registration group is up to full speed it can issue URLs that can be used with any DAV compliant system, new or old. 2) Why don't you use a URN namespace of the form URN:DAV:xxxx? People who know me are especially likely to ask this question as I am an unabashed fan of URNs. The reason for avoiding the URN namespace is that the whole space is still being defined. For example, they only have "a very drafty draft" [ftp://ietf.org/internet-drafts/draft-ietf-urn-nid-req-02.txt] on how URN namespaces are registered. Additionally we are not proposing any mechanism for actually resolving the DAV URI namespace so it seems inappropriate to define a URN namespace with no provision for resolution. There is nothing that stops us from coming back later, defining a URN:DAV namespace and putting that value in href. 3) Wait, does this mean I can't name my XML elements or DAV properties with arbitrary URIs? ABSOLUTELY NOT. DAV is 100% compliant with the XML namespace extension which allows for ANY LEGAL URI to be used to name either an XML element or a property. As such you can use ANY URI you want. Thus, for example, the Dublin group could define a new author property with a namespace declaration of <?namespace name=http://www.aiim.org/dublin/standards/properties/author as="dublin">. This proposal ONLY effects the names DAV uses for the XML elements and properties it defines. All DAV needs for its own XML elements and properties is uniqueness, not resolution. As such I propose we stick with the feature we need rather than trying to go beyond to features that are certainly sexy but not necessary. Yaron
Received on Friday, 30 January 1998 19:07:24 UTC