Extensions for Distributed Authoring and Versioning on the World Wide Web -- WEBDAV

It would seem best to implement a flat property namespace, inducing a natural isomorphism between DAV and most native file systems. Adopting such a model will not restrict RDBS from taking full advantage of their search facilities.

However, it seems that future trends might be toward hierarchical properties. Therefore, DAV requirements [Slein et al.] stipulate that the design of the flat property system MUST be such that it will be possible to add true hierarchical properties later without breaking downlevel clients. Specifically, a flat client MUST be able to speak to a hierarchical server and a hierarchical client MUST be able to speak to a flat server. Worst case either way MUST be that the request fails.

2.2.2.3 Property Names

The URI framework provides a mechanism to prevent namespace collision and for varying degrees of administrative control. Rather than reinvent these desirable features, DAV properties make use of them by requiring that all DAV property names MUST be URIs. Since a property is also an XML element, the name of the XML element is a URI.

The property namespace is flat, that is, it is not possible to string together a series of property names in order to refer to a hierarchy of properties. Thus it is possible to refer to a property B but not a property A/B, where A is also a property defined on the resource.

Finally, it is not possible to define the same property twice as this would cause a collision in the resource's property namespace.

2.3 Schemas

Schema discovery is used to determine if a system supports a group of properties or XML elements. A property does not necessarily contain sufficient information to identify any schema(s) to which it may belong.

As with property names, schemas MUST use URIs as their names.

A resource declares its support for a schema by defining a property whose name is the same as the schema's. The property SHOULD contain the PropSchema XML element.

2.3.1 PropSchema XML Element

2.3.2 DTD XML Element

2.3.3 DefinedProps XML Element

2.3.4 PropEntries XML Element

2.3.5 Live XML Element

2.4 DAV Schema

2.4.1 DAV Property

2.4.2 Level XML Element

2.4.3 Prop XML element

2.4.4 PropLoc XML Attribute

2.4.5 Example

<?XML:Namespace href="http://www.ietf.org/standards/dav/" AS="D"/>
<?XML:Namespace href="AIIM:Dublin:" AS="A"/> <D:Prop>
     <A:Author D:PropLoc="http://www.foo.com/resource/props/Author">
          Larry Masinter
     </A:Author>
</D:Prop>

2.5 Property Identifiers

2.5.1 Problem Definition

A server is free to assign whatever URI it chooses to identify an instance of a property defined on a resource. In fact, a server is free not to reveal the URI of an instance of a particular resource and instead require that the client access the property through PROPFIND and PROPPATCH. However, many servers will want to allow clients to directly manipulate properties. On these servers, a client can discover the URI of an instance of a property by performing a PROPFIND and examining the PropLoc attribute, if returned, of each property.

2.6 Link XML Element

2.6.1 Problem Description

2.6.2 Solution Requirements

2.6.3 Link XML Element

2.6.4 Src XML Element

2.6.5 Dst XML Element

2.6.6 Example

<?XML:Namespace
     href = "http://www.ietf.org/standards/dav/" AS = "D"/>
<?XML:Namespace
     href = "http://www.foocorp.com/Project/" AS = "F"/>
<D:Prop>
     <Source>
          <Link>
               <F:ProjFiles>Source</F:ProjFiles>
               <src>http://foo.bar/program</src>
               <dst>http://foo.bar/src/main.c</dst>
          </Link>
          <Link>
               <F:ProjFiles>Library</F:ProjFiles>
               <src>http://foo.bar/program</src>
               <dst>http://foo.bar/src/main.lib</dst>
          </Link>
          <Link>
               <F:ProjFiles>Makefile</F:ProjFiles>
               <src>http://foo.bar/program</src>
               <dst>http://foo.bar/src/makefile</dst>
     <Link>
     </Source>
</D:Prop>

2.7 Multi-Status Response

2.7.1 Problem Definition

2.7.2 Solution Requirements

2.7.3 Multi-Status Response

2.7.3.1 Multiresponse

2.7.3.2 Response

2.7.3.3 Status

2.7.3.4 ResponseDescription

2.8 Properties and Methods

2.8.1 DELETE

2.8.2 GET

2.8.2.1 Example

GET /bar.html;prop=z39.50_authors HTTP/1.1
Host: foo.com

HTTP/1.1 200 OK
Content-Type: text/xml
Content-Length: xxxx
E-tag: "1234"
Last-Modified: xxxx

<?XML:Namespace href = "http://www.ietf.org/standards/dav/" AS = "D"/>
<?XML:Namespace href = "http://www.w3.com/standards/z39.50/"AS = "Z"/>
<D:prop>
     <Z:Authors>
          <Z:Author>Jane Doe</Z:Author>
          <Z:Author>Joe Doe</Z:Author>
           <Z:Author>Lots o'Doe</Z:Author>
     </Z:Authors>
</D:prop>

2.8.3 PROPPATCH

All DAV compliant servers MUST process instructions which are specified using the PropertyUpdate, Create, and Remove XML elements of the DAV schema. The request message body MUST contain at least one PropertyUpdate XML element. Instruction processing MUST occur in the order instructions are received (i.e., from top to bottom), and MUST be performed atomically.

2.8.3.1 PropertyUpdate XML element

2.8.3.2 Create XML element

2.8.3.3 Remove XML element

2.8.3.4 Response

2.8.3.5 Response Codes

2.8.3.6 Example

PROPPATCH /bar.html HTTP/1.1
Host: www.foo.com
Content-Type: text/xml
Content-Length: xxxx

<?XML:Namespace href = "http://www.ietf.org/standards/dav/" AS = "D"/>
<?XML:Namespace href = "http://www.w3.com/standards/z39.50/" AS = "Z"/>
<D:PropertyUpdate>
     <Create>
          <prop>
               <Z:authors>
                    <Z:Author>Jim Whitehead</Z:Author>
                    <Z:Author>Roy Fielding</Z:Author>
               </Z:authors>
          </Prop>
     </Create>
     <Remove>
          <prop><Z:Copyright-Owner/></prop>
     </Remove>
</D:PropertyUpdate>
HTTP/1.1 405 Conflict
Content-Type: text/xml
Content-Length: xxxxx

<?XML:Namespace href="http://www.ietf.org/standards/dav/" AS = "D"/>
<?XML:Namespace href="http://www.w3.com/standards/z39.50/" AS = "Z"/>
<D:MultiResponse>
     <ResponseDescription> Copyright Owner can not be deleted or
altered.</ResponseDescription>
     <Response>
          <Prop><Z:authors/></Prop>
          <Status>HTTP/1.1 418 Atomicity Failure</Status>
     </Response>
     <Response>
          <Prop><Z:Copyright-Owner/></Prop>
          <Status>HTTP/1.1 405 Conflict</Status>
     </Response>
</D:MultiResponse>

2.8.4 PUT

2.8.5 PROPFIND

The response is a text/xml message body that contains a Multiresponse XML element which describes the results of the attempts to retrieve the various properties. If a property was successfully retrieved then its value MUST be returned in the Prop XML element. In the case of Allprop and Findprop, if a principal does not have the right to know if a particular property exists, an error MUST NOT be returned. The results of this method SHOULD NOT be cached.

2.8.5.1 PropFind XML element

2.8.5.2 Allprop

2.8.5.3 Propname

2.8.5.4 PropFindResult XML element

2.8.5.5 Example 1 - Prop

PROPFIND  /container/ HTTP/1.1
Host: www.foo.bar
Content-Length: xxxx
Content-Type: text/xml

<?XML:Namespace href =
     "http://www.ietf.org/standards/dav/" AS = "G"/>
<?XML:Namespace href =
     "http://www.foo.bar/boxschema/" AS = "B"/>
<G:PROPFIND>
     <prop>
          <B:bigbox>
          <B:author>
          <B:DingALing>
          <B:Random>
     </prop>
</G:PROPFIND>

HTTP/1.1 207 Multi-Response
Content-Type: text/xml
Content-Length: xxxxx

<?XML:Namespace href ="http://www.ietf.org/standards/dav/" AS = "S">
<?XML:Namespace href = "http://www.foo.bar/boxschema" AS = R">
<D:MultiResponse>
     <ResponseDescription> There has been an access violation
error. </ResponseDescription>
     <Response>
          <Prop>
               <R:bigbox D:Proploc="http://prop.com/BoxType">
                    <BoxType>Box type A</BoxType>
               </R:bigbox>
               <R:author D:Proploc="http://prop.com/Author">
                    <Name>J.J. Dingleheimerschmidt</Name>
               </R:author>
          </Prop>
          <Status>HTTP/1.1 200 Success</Status>
</Response>
     <Response>
          <Prop><R:DingALing/><R:Random/></>
          <Status>HTTP/1.1 403 Forbidden</Status>
          <ResponseDescription> The user does not have access to
the DingALink property. </ResponseDescription>
     </Response>
</D:MultiResponse>

2.8.5.6 Example 2 - Allprop

PROPFIND  /container/ HTTP/1.1
Host: www.foo.bar
Content-Length: xxxx
Content-Type: text/xml

<?XML:Namespace href =
     "http://www.ietf.org/standards/dav/" AS = "G"/>
<G:PROPFIND>
     <Allprop/>
</G:PROPFIND>

HTTP/1.1 200 Success
Content-Type: text/xml
Content-Length: xxxxx

<?XML:Namespace href =
     "http://www.ietf.org/standards/dav/" As = "S">
<?XML:Namespace href = "http://www.foo.bar/boxschema" AS = R">
<S:MultiResponse>
     <Prop>
          <R:bigbox D:Proploc="http://prop.com/BigBox">
               <BoxType>Box type A</BoxType>
          </R:bigbox>
          <R:author D:Proploc="http://prop.com/Author">
               <Name>Hadrian</Name>
          </R:author>
     </Prop>
     <Status>HTTP/1.1 200 Success</Status>
</S:MultiResponse>

2.8.5.7 Example 3 - Propname

PROPFIND  /container/ HTTP/1.1
Host: www.foo.bar
Content-Length: xxxx
Content-Type: text/xml

<?XML:Namespace href = "http://www.ietf.org/standards/dav/" AS = "G"/>
<G:PROPFIND>
     <Propname/>
</G:PROPFIND>

HTTP/1.1 200 Success
Content-Type: text/xml
Content-Length: xxxxx

<?XML:Namespace href = "http://www.ietf.org/standards/dav/" As = "S">
<?XML:Namespace href = "http://www.foo.bar/boxschema" AS = "R">
<S:MultiResponse>
     <Prop>
          <R:bigbox D:Proploc="http://prop.com/BigBox"/>
          <R:author D:Proploc="http://prop.com/Author"/>
          <R:DingALing/>
          <R:Random/>
     </Prop>
     <Status>HTTP/1.1 200 Success</Status>
</S:MultiResponse>

3 A Proposal for Collections of Web Resources and Name Space Operations

3.1 Observations on the HTTP Object Model

3.1.1 Collection Resources

The purpose of a collection resource is to model collection-like objects (e.g., a filesystem directory) within a server's namespace. Once these objects have been modeled with collections, a client may perform an INDEX, add and remove external members using ADDREF and DELREF, and perform recursive operations, such as a full hierarchy copy.

To support methods which operate on collections, a server SHOULD model its collection-like objects with collection resources. For example, a server which is implemented on top of a filesystem SHOULD treat all filesystem directories exposed by the server as collection resources.

3.1.2 Creation and Retrieval of Collection Resources

In HTTP/1.1, the PUT method is defined to store the request body at the location specified by Request-URI. While a description format for a collection can readily be constructed that could be used with PUT, the implications of sending such a description to the server are undesirable. For example, if a description of a collection that omitted some existing resources were PUT to a server, this might be interpreted as a command to remove those members. This would extend PUT to perform DELETE functionality, which is undesirable since it changes the semantics of PUT, and makes it difficult to control DELETE functionality with an access control scheme based on methods.

While the POST method is sufficiently open-ended that a "create a collection" POST command could be constructed, this is undesirable because it would be difficult to separate access control for collection creation from other uses of POST if they both use the same method.

While it might seem desirable to have GET return a listing of the members of a collection, this is foiled by the existence of the "index.html" de-facto standard namespace redirection, in which a GET request on a collection is automatically redirected to the index.html resource.

The exact definition of the behavior of GET and PUT on collections is defined later in this document.

3.1.2.1 Example

If http://foo/bar/b/d is created with a MKCOL and http://foo/bar/b/d/e was created then a DELETE on d would fail because d is a collection with an internal member. However the existence of the collection d is something of an illusion. If a DELETE was executed on http://foo/bar then everything would be deleted, even though http://foo/bar/b/d was created with a MKCOL.

Thus the effect of a MKCOL within a composite resource's namespace is felt on its children, not its ancestors. The children of d MUST be treated as members of a collection when a method is executed on d. But a method executed on b or a is treated as if there only existed a non-collection resource.

3.1.3 Source Resources and Output Resources

However, the server can sometimes process HTML resources before they are transmitted as a return entity body. For example, server-side-include directives within an HTML file instruct a server to replace the directive with another value, such as the current date. In this case, what is returned by GET (HTML plus date) differs from the persistent state of the resource (HTML plus directive). Typically there is no way to access the HTML resource containing the unprocessed directive.

Sometimes the entity returned by GET is the output of a data-producing process that is described by one or more source resources (that may not even have a location in the URL namespace). A single data-producing process may dynamically generate the state of a potentially large number of output resources. An example of this is a CGI script that describes a "finger" gateway process that maps part of the namespace of a server into finger requests, such as http://www.foo.bar.org/finger_gateway/user@host.

In the absence of distributed authoring capability, it is acceptable to have no mapping of source resource(s) to the URI namespace, and in fact has desirable security benefits. However, if remote editing of the source resource(s) is desired, the source resource(s) should be given a location in the URI namespace. This source location should not be one of the locations at which the generated output is retrievable, since in general it is impossible for the server to differentiate requests for source resources from requests for process output resources. There is often a many-to-many relationship between source resources and output resources.

For DAV compliant servers all output resources which have a single source resource (and that source resource has a URI), the URI of the source resource SHOULD be stored in a single link on the output resource with type http://www.ietf.org/standards/dav/source. Note that by storing the source URI in links on the output resources, the burden of discovering the source is placed on the authoring client.

3.2 MKCOL Method

3.2.1 Problem Description

3.2.2 Solution Requirements

3.2.3 Request

3.2.3.1 MKCOL Without Request Body

3.2.3.2 MKCOL With Request Body

3.2.4 Response

3.2.5 Example

     MKCOL /webdisc/xfiles/ HTTP/1.1
     Host: www.server.org

     HTTP/1.1 201 Created

3.3 Standard DAV Properties

3.3.1 IsCollection Property

3.3.2 DisplayName Property

3.3.3 CreationDate Property

3.3.4 GETentity Property

3.3.5 INDEXentity Property

3.3.6 Content-Type XML Element

3.3.7 Content-Length XML Element

3.3.8 Content-Language XML Element

3.3.9 Last-Modified XML Element

3.3.10 Etag XML Element

3.4 INDEX Method

3.4.1 Problem Description

3.4.2 Solution Requirements

3.4.3 The Request

For a collection, INDEX MUST return a list of its members. All WebDAV compliant resources MUST support the text/xml response entity described below. The INDEX result for a collection MAY also return a list of the members of child collections, to any depth.

Collections that respond to an INDEX method with a text/xml entity MUST contain only one ResInfo element. This ResInfo element contains an Href element, which gives the identifier(s) of the resource, a Prop element, which gives selected properties of the resource, and a Members element, which contains a ResInfo element for each member of the collection. The Prop element MUST contain at least the following properties, if they are defined and available: DisplayName, IsCollection, CreationDate, GETentity, and INDEXentity.

The response from INDEX is cacheable, and SHOULD be accompanied by an ETag header (see section 13.3.4 of RFC 2068). If GET and INDEX return different entities for the same resource state, they MUST return different entity tags.

3.4.4 The Response

3.4.5 ResInfo XML Element

3.4.6 Members XML Element

3.4.7 Href XML Element

3.4.8 Example

INDEX /user/yarong/dav_drafts/ HTTP/1.1
Host: www.microsoft.com

HTTP/1.1 200 OK
Content-Type: text/xml
Content-Length: xxx
Last-Modified: Thu, 11 Sep 1997 23:45:12 GMT
ETag: "fooyyybar"

     <?XML:Namespace href = "http://www.ietf.org/standards/dav/" as = "D"/>
<D:ResInfo>
     <XML:Href>
          http://www.microsoft.com/user/yarong/dav_drafts/
     </XML:Href>
     <Prop>
        <DisplayName>
          WebDAV working drafts directory
        </DisplayName>
        <IsCollection>true</IsCollection>
        <CreationDate>19970418T070304Z</CreationDate>
        <GETentity>
          <Content-Type>text/html</Content-Type>
          <Content-Length>2754</Content-Length>
          <Content-Language>en</Content-Language>
          <Last-Modified>
               Fri, 22 Aug 1997 10:11:26 GMT
          </Last-Modified>
          <Etag>"8675309"</Etag>
        </GETentity>
        <INDEXentity>
          <Content-Type>text/xml</Content-Type>
          <Content-Length>xxx</Content-Length>
          <Last-Modified>
               Thu, 11 Sep 1997 23:45:12 GMT
          </Last-Modified>
          <Etag>"fooyyybar"</Etag>
        </INDEXentity>
     </Prop>

     <Members>
        <ResInfo>
          <XML:Href>
          http://www.microsoft.com/user/yarong/dav_drafts/base
          </XML:Href>
          <Prop>
             <IsCollection
     D:PropLoc="http://www.microsoft.com/user/yarong/dav_drafts/b
     ase;props=IsCollection">
               False
             </IsCollection>
             <DisplayName>
               WebDAV Name Space Operations Draft
             </DisplayName>
             <Creation-Date>19970320T230525Z</Creation-Date>

             <GETentity>
               <Content-Type>application/msword</Content-Type>
               <Content-Length>1400</Content-Length>
               <Content-Language>en</Content-Language>
               <Last-Modified>
                    Fri, 22 Aug 1997 18:22:56 GMT
               </Last-Modified>
               <Etag>"8675309"</Etag>
             </GETentity>
          </Prop>
        </ResInfo>
     </Members>
     </D:ResInfo>

3.5 Behavior of RFC 2068 Methods on Collections

3.5.1 GET, HEAD for Collections

Similarly, since the definition of HEAD is a GET without a response message body, the semantics of HEAD are unmodified when applied to collection resources.

3.5.2 POST for Collections

3.5.3 PUT for Collections

When the PUT operation creates a new non-collection resource all ancestors MUST already exist. If all ancestors do not exist, the method MUST fail with a 409 Conflict status code. For example, if resource /a/b/c/d.html is to be created and /a/b/c/ does not exist, then the request MUST fail.

3.5.3.1 PUT for Non-Collection Resources

3.5.4 DELETE for Collections

3.5.5 DELETE Method for Non-Collection Resources

3.6 COPY Method

3.6.1 Problem Description

This is problematic because of the network traffic involved in making a copy, and because there is often no way to fully express a resource as an entity without a loss of fidelity.

3.6.2 Solution Requirements

3.6.3 The Request

3.6.3.1 COPY for HTTP/1.1 resources

All properties on the source resource MUST be duplicated on the destination resource, subject to modifying headers, following the definition for copying properties.

3.6.3.2 COPY for Properties

If a property on the source already exists on the resource and the overwrite header is set to TRUE then the property at the destination MUST be overwritten with the property from the source. If the overwrite header is false and the previous situation exists then the COPY MUST fail with a 409 Conflict.

3.6.3.3 COPY for Collections

3.6.3.4 Type Interactions

If the destination resource identifies a collection and the Overwrite header is "true," prior to performing the copy, the server MUST perform a DELETE operation on the collection.

3.6.4 The Response

3.6.5 Examples

3.6.5.1 Overwrite Example

COPY /~fielding/index.html HTTP/1.1
Host: www.ics.uci.edu
Destination: http://www.ics.uci.edu/users/f/fielding/index.html

HTTP/1.1 200 OK

3.6.5.2 No Overwrite Example

COPY /~fielding/index.html HTTP/1.1
Host: www.ics.uci.edu
Destination: http://www.ics.uci.edu/users/f/fielding/index.html
Overwrite: "false"

HTTP/1.1 412 Precondition Failed

3.7 MOVE Method

3.7.1 Problem Description

With a WEBDAV server, a principal may accomplish this task by issuing a COPY and then DELETE. Network load decreases, but the server load may still be significant because the server must create a duplicate resource. Were a server to know beforehand that a principal intended to perform COPY and DELETE operations in succession, it could avoid the creation of a duplicate resource.

3.7.2 Solution Requirements

3.7.3 The Request

3.7.4 The Response

3.7.5 Examples

3.7.5.1 Overwrite Example

MOVE /~fielding/index.html HTTP/1.1
Host: www.ics.uci.edu
Destination: http://www.ics.uci.edu/users/f/fielding/index.html

HTTP/1.1 200 OK
Content-Location:
http://www.ics.uci.edu/users/f/fielding/index.html

3.8 ADDREF Method

3.8.1 Problem Definition

3.8.2 Solution Requirements

3.8.3 The Request

It is not an error if the URI specified in the Collection-Member header already exists as an external member of the collection, however, after processing ADDREF there MUST be only one instance of the URI in the collection. If the URI specified in the Collection-Member header already exists as an internal member of the collection, the ADDREF method MUST fail with a 412 Precondition Failed status code.

3.8.4 Example

ADDREF /~whitehead/dav/ HTTP/1.1
HOST: www.ics.udi.edu
Collection-Member: http://www.ietf.org/standards/dav/

HTTP/1.1 200 OK

3.9 DELREF Method

3.9.1 Problem Definition

3.9.2 Solution Requirements

3.9.3 The Request

3.9.4 Example

DELREF /~whitehead/dav/ HTTP/1.1
Host: www.ics.udi.edu
Collection-Member: http://www.ietf.org/standards/dav/

HTTP/1.1 200 OK

3.10 PATCH Method

3.10.1 Problem Definition

3.10.2 Solution Requirements

3.10.3 The Request

If the request appears (at least initially) to be acceptable, the server MUST transmit an interim 100 response message after receiving the empty line terminating the request headers and continue processing the request. Since the semantics of PATCH are non-idempotent, responses to this method are not cacheable.

While server support for PATCH is optional, if a server does support PATCH, it MUST support at least the text/xml diff format defined below. Support for the VTML difference format [VTML] is recommended, but not required.

3.10.4 text/xml elements for PATCH

The changes specified by the resourceupdate XML element MUST be executed atomically.

3.10.4.1 ResourceUpdate

3.10.4.2 Insert

Please note that in order to protect the white space contained in this XML element the following attribute/value MUST be included in the element: XML-SPACE = "PRESERVE".

3.10.4.3 Delete

3.10.4.4 Replace

Please note that in order to protect the white space contained in this XML element the following attribute/value MUST be included in the element: XML-SPACE = "PRESERVE".

3.10.4.5 Octet-Range Attribute

3.10.5 The Response

3.10.6 Examples

3.10.6.1 HTML file modification

     <HTML>
     <HEAD>
     <TITLE>Hello world HTML page</TITLE>
     </HEAD>
     <BODY>
     <P>Hello, world!</P>
     </BODY>
     </HTML>

     PATCH hello.html HTTP/1.1
     Host: www.example.org
     Content-Type: text/xml
     Content-Length: xxx

                                   HTTP/1.1 100 Continue
     <?XML:Namespace href =
     Shttp://www.ietf.org/standards/dav/patch/" AS = "D"/>
     <D:ResourceUpdate>
          <Replace XML-SPACE = "PRESERVE">
          <octet-range>14</octet-range>
          &003CTITLE&003ENew Title&003C/TITLE&003E
          </Replace>
          <Delete><octet-range>38-50</Delete>
     <Insert XML-SPACE = "PRESERVE"><octet-range>86</>&
     003CP&003ENew paragraph&003C/P&003E</Insert>
     </D:ResourceUpdate>
                                   HTTP/1.1 200 OK

     <HTML>
     <HEAD>
     <TITLE>New Title</TITLE>
     </HEAD>
     <BODY>
     <P>Hello, world!</P>
     <P>New paragraph</P>
     </BODY>
     </HTML>

3.11 Headers

3.11.1 Destination Header

     Destination= "Destination" ":" URI

3.11.2 Enforce-Live-Properties Header

     EnforceLiveProperties = "Enforce-Live-Properties" ":" ("*" |
     "Omit" | 1#(Property-Name))
     Property-Name = "<" URI ">"

3.11.3 Overwrite Header

     Overwrite = "Overwrite" ":" ("true" | "false")

3.11.4 Destroy Header

The NoUndelete token requests that the resource MUST NOT be left in a state such that it can be undeleted.

The VersionDestroy token includes the functionality of the NoUndelete token and extends it to include having the server remove all versioning references to the resource that it has control over.

     DestroyHeader = "Destroy" ":" #Choices

3.11.5 Collection-Member Header

     CollectionMember = "Collection-Member" ":" URI

3.12 Links

3.12.1 Source Link Property Type

4 State Tokens

4.1 Overview

4.1.1 Problem Description

The fundamental issue with entity tags is that they can only be generated by a resource. However there are times when a client will want to be able to share state tokens between resources, potentially on different servers, as well as be able to generate certain types of lock tokens without first having to communicate with a server.

For example, a principal may wish to require that resource B have a certain state in order for a method to successfully execute on resource A. If the client submits an e-tag from resource B to resource A, then A has no way of knowing that the e-tag is meant to describe resource B.

Another example occurs when a principal wishes to predicate the successful completion of a method on the absence of any locks on a resource. It is not sufficient to submit an "If-None-Match: *" as this refers to the existence of an entity, not of a lock.

This draft defines the term "state token" as an identifier for a state of a resource. The sections below define requirements for state tokens and provide a state token syntax, along with two new headers which can accept the new state token syntax.

4.1.2 Solution Requirements

4.1.2.1 Syntax

This self-describing nature allows servers to accept tokens from other servers and potentially be able to coordinate state information cross resource and cross site through standardized protocols. For example, the execution of a request on resource A can be predicated on the state of resource B, where A and B are potentially on different servers.

Client Generable. The state token syntax must allow, when appropriate, for clients to generate a state token without having first communicated with a server.

One drawback of entity tags is that they are set by the server, and there is no interoperable algorithm for calculating an entity tag. Consequently, a client cannot generate an entity tag from a particular state of a resource. However, a state token which encodes an MD5 state hash could be calculated by a client based on a client-held state of a resource, and then submitted to a server in a conditional method invocation.

Another potential use for client generable state tokens is for a client to generate lock tokens with wild card fields, and hence be able to express conditionals such as: "only execute this GET if there are no write locks on this resource."

4.1.2.2 Conditonals

4.2 State Token Syntax

State-Token = "StateToken:" Type ":" Resources ":" State-Info
Type = "Type" "=" Caret-encoded-URL
Resources = "Res" "=" Caret-encoded-URL
Caret-encoded-URL = "^" Resource "^"
Resource = <A URI where all "^" characters are escaped>
State-Info = *(uchar | reserved)  ; uchar, reserved defined
section 3.2.1 of RFC 2068

This proposal has created a new URL scheme for state tokens because a state token names a network resource using its normal name, which is typically state-invariant, along with additional information that specifies a particular state of the resource. Encoding the state information into the native URL scheme of the network resource was not felt to be safe, since freedom from name space collisions could not be guaranteed. If this proposal is accepted, the StateToken URL scheme will need to be defined and registered with IANA.

State Token URLs begin with the URL scheme name "StateToken" rather than the name of the particular state token type they represent in order to make the URL self describing. Thus it is possible to examine the URL and know, at a minimum, that it is a state token.

Labeled name/value pairs are used within the token to allow new fields to be added. Processors of state tokens MUST be prepared to accept the fields in whatever order they are present and MUST ignore any fields they do not understand. The "Type" field specifies the type of the state information encoded in the state token. A URL is used in order to avoid namespace collisions.

The "Res" field identifies the resource for which the state token specifies a particular state. Since commas and spaces are acceptable URL characters, a caret is used to delimit a URL. Since a caret is an acceptable URL character, any instances of it must be escaped using the % escape convention.

The State-Info production is expanded upon in descriptions of specific state token types, and is intended to contain the state description information for a particular state token.

4.3 State Token Conditional Headers

4.3.1 If-State-Match

If-State-Match = "If-State-Match" ":" ("AND" | "OR") 1#("<" State-Token ">")

The If-State-Match header is intended to have similar functionality to the If-Match header defined in section 14.25 of RFC 2068.

If the AND keyword is used and all of the state tokens identify the state of the resource, then the server MAY perform the requested method. If the OR keyword is used and any of the state tokens identifies the current state of the resource, then server MAY perform the requested method. If neither of the keyword requirements is met, the server MUST NOT perform the requested method, and MUST return a 412 (Precondition Failed) response.

4.3.2 If-None-State-Match

If-None-State-Match = "If-None-State-Match" ":" 1#("<" State-Token ">")

If any of the state tokens identifies the current state of the resource, the server MUST NOT perform the requested method. Instead, if the request method was GET, HEAD, INDEX, or GETMETA, the server SHOULD respond with a 304 (Not Modified) response, including the cache-related entity-header fields (particularly ETag) of the current state of the resource. For all other request methods, the server MUST respond with a status of 412 (Precondition Failed).

If none of the state tokens identifies the current state of the resource, the server MAY perform the requested method.

Note that the "AND" and "OR" keywords specified with the If-State-Match header are intentionally not defined for If-None-State-Match, because this functionality is not required.

4.4 State Token Header

State-Token-Header = "State-Token" ":" 1#("<" State-Token ">")

4.5 E-Tags

5 Locking

5.1 Locking: Introduction

This specification allows locks to vary over two parameters, the number of principals involved and the type of access to be granted. Furthermore, this document only provides the definition of locking for one access type, write. However, the syntax is extensible, and allows the specification of other access types.

5.1.1 Exclusive Vs. Shared Locks

However, there are times when the goal of a lock is not to exclude others from exercising an access right but rather to provide a mechanism for principals to indicate that they intend to exercise their access right. Shared locks are provided for this case. A shared lock allows multiple principals to receive a lock, hence any principal with appropriate access can get the lock.

With shared locks there are two trust sets that affect a resource. The first trust set is created by access permissions. Principals who are trusted, for example, may have permission to write the resource, those who are not, don't. Among those who have access permission to write the resource, the set of principals who have taken out a shared lock also must trust each other, creating a (typically) smaller trust set within the access permission write set.

Starting with every possible principal on the Internet, in most situations the vast majority of these principals will not have write access to a given resource. Of the small number who do have write access, some principals may decide to guarantee their edits are free from overwrite conflicts by using exclusive write locks. Others may decide they trust their collaborators (the potential set of collaborators being the set of principals who have write permission) and use a shared lock, which informs their collaborators that a principal is potentially working on the resource.

The WebDAV extensions to HTTP do not need to provide all of the communications paths necessary for principals to coordinate their activities. When using shared locks, principals may use any out of band communication channel to coordinate their work (e.g., face-to-face interaction, written notes, post-it notes on the screen, telephone conversation, email, etc.) The intent of a shared lock is to let collaborators know who else is potentially working on a resource.

Shared locks are included because experience from web distributed authoring systems has indicated that exclusive write locks are often too rigid. An exclusive write lock is used to enforce a particular editing process: take out exclusive write lock, read the resource, perform edits, write the resource, release the lock. What happens if the lock isn't released? While the time-out mechanism provides one solution, if you need to force the release of a lock immediately, it doesn't help much. Granted, an administrator can release the lock for you, but this could become a significant burden for large sites. In addition there is the problem that an administrator may not be immediately available.

Despite their potential problems, exclusive write locks are extremely useful, since often a guarantee of freedom from overwrite conflicts is what is needed. The tradeoff described in this specification is to provide exclusive write locks, but also to provide a less strict mechanism in the form of shared locks which can be used by a set of people who trust each other and who have access to a communications channel external to HTTP which can be used to negotiate writing to the resource.

5.1.2 Required Support

The reason for this flexibility is that server implementers have said that they are willing to accept minimum requirements on all services but locking. Locking policy strikes to the very heart of their resource management and versioning systems and they require control over what sort of locking will be made available. For example, some systems only support shared write locks while others only provide support for exclusive write locks while yet others use no locking at all. As each system is sufficiently different to merit exclusion of certain locking features, the authors are proposing that locking be allowed as the sole axis of negotiation within WebDAV.

5.2 LOCK Method

5.2.1 Operation

A successful response to a lock invocation MUST include Lock-Token and Time-Out headers.

5.2.2 The Effect of Locks on Properties and Containers

For containers, a lock also affects the ability to add or remove members. The nature of the effect depends upon the type of access control involved.

5.2.3 Locking Replicated Resources

5.2.4 Locking Multiple Resources

The Lock-Tree option of the lock request specifies that the resource and all its internal children (including internal collections, and their internal members) are to be locked. This is another mechanism by which a request for a lock on multiple resources can be specified.

Currently existing locks can not be extended to cover more or less resources, and any request to expand or contract the number of resources in a lock MUST fail with a 409 Conflict status code. So, for example, if resource A is exclusively write locked and then the same principal asked to exclusively write lock resources A, B, and C, the request would fail as A is already locked and the lock can not be extended.

A successful result will return a single lock token which represents all the resources that have been locked. If an UNLOCK is executed on this token, all associated resources are unlocked.

If the lock can not be granted to all resources, a 406 Conflict status code MUST be returned with a response entity body containing a multiresponse XML element describing which resource(s) prevented the lock from being granted.

5.2.5 Interaction with other Methods

5.2.6 Lock Compatibility Table

Legend: True = lock MAY be granted. False = lock MUST NOT be granted. *=if the principal requesting the lock is the owner of the lock, the lock MAY be regranted.

The current lock state of a resource is given in the leftmost column, and lock requests are listed in the first row. The intersection of a row and column gives the result of a lock request. For example, if a shared lock is held on a resource, and an exclusive lock is requested, the table entry is "false", indicating the lock must not be granted.

If an exclusive or shared lock is re-requested by the principal who owns the lock, the lock MUST be regranted. If the lock is regranted, the same lock token that was previously issued MUST be returned.

5.2.7 Status Codes

5.2.8 Lock-Info Request Header

LockInfo = "Lock-Info" ":" DAVLockType SP DAVLockScope [SP
AdditionalLocks] [SP Lock-Tree]
DAVLockType = "LockType" "=" DAVLockTypeValue
DAVLockTypeValue = ("Write" | *(uchar | reserved))
DAVLockScope = "LockScope" "=" DAVLockScopeValue
DAVLockScopeValue = ("Exclusive" |"Shared" | *(uchar | reserved))
AdditionalLocks = "AddLocks" "=" 1*("<" URI ">")
Lock-Tree = "Lock-Tree" "=" ("True" | "False")

5.2.9 Owner Request Header

5.2.9.1 Problem Description

5.2.9.2 Solution Requirements

The From header (defined in RFC 2068), which contains only an email mailbox, is not sufficient for the purposes of quick identification. When desperately looking for someone to remove a lock, e-mail is often not sufficient. A telephone number (cell number, pager number, etc.) would be better. Furthermore, the email address in the From header only optionally includes the owners name and that name is often set to an alias anyway. Therefore a header more flexible than From is required.

The value also needs to be such that both man and machine can place values in it and later retrieve those values.

5.2.9.3 Syntax

Owner = "Owner" ":" (Coded-XML  | quoted-string)
Coded-XML = field-content    ; XML where any character which is
  not legal in field-content (see section 4.2 of [Fielding et al.,
  1997]) is XML encoded

5.2.10 Time-Out Header

5.2.10.1 Problem Description

5.2.10.2 Solution Requirements

User-agents MUST assume that locks may arbitrarily disappear at any time. If their actions require confirmation of the existence of a lock then the If-State headers are available.

5.2.10.3 Syntax

TimeOut = "Time-Out" ":" 1#TimeType
TimeType = ("Second-" DAVTimeOutVal | "Infinite" | Extend)
DAVTimeOutVal = 1*digit
Extend = RFC-Reg | URL "-" Token ; The URL format is used for
unregistered TimeTypes
RFC-Req = Token ; This is a TimeType that has been published as an RFC

A Time-Out request header MUST contain at least one TimeType and MAY contain multiple TimeType entries. The purpose of listing multiple TimeType is to indicate multiple different values and value types that are acceptable to the client. The client lists the TimeType entries in order of preference.

The Time-Out response header MUST use a Second value, Infinite, or a TimeType the client has indicated familiarity with. The server MAY assume a client is familiar with any TimeType submitted in a Time-Out header.

The "Second" TimeType specifies the number of seconds that MUST elapse between granting of the lock at the server, and the automatic removal of the lock. A server MUST not generate a time out value for "Second" greater than 2^32-1.

The time out counter is restarted any time the client sends a method to any member of the lock, including unsupported methods, or methods which are unsuccessful. It is recommended that the HEAD method be used when the goal is simply to restart the time out counter.

If the timeout expires then the lock is lost. Specifically the server SHOULD act as if an UNLOCK method was executed by the server on the resource using the lock token of the timed-out lock, performed with its override authority. Thus logs, notifications, and other mechanisms that act as side effects to the granting and removal of a lock will be properly informed as to the disposition of the lock.

Servers are advised to pay close attention to the values submitted by clients, as they will be indicative of the type of activity the client intends to perform. For example, an applet running in a browser may need to lock a resource, but because of the instability of the environment within which the applet is running, the applet may be turned off without warning. As a result, the applet is likely to ask for a relatively small time-out value so that if the applet dies, the lock can be quickly harvested. However a document management system is likely to ask for an extremely long time-out because its user may be planning on going off-line.

5.2.11 Lock Response

5.2.11.1 Lock-Token Response Header

Lock-Token = "Lock-Token" ":" URI

5.2.12 Example - Simple Lock Request

LOCK /workspace/webdav/proposal.doc HTTP/1.1
Host: webdav.sb.aol.com
Lock-Info: LockType=Write LockScope=Exclusive
Time-Out: Infinite; Second-4100000000
Owner: <?XML:Namespace href="http://www.ietf.org/standards/dav/" AS =
"D"/>
 <D:HREF>http://www.ics.uci.edu/~ejw/contact.html</D:HREF>

HTTP/1.1 200 OK
Lock-Token: OpaqueLockToken:xyz122393481230912asdfa09s8df09s7df08
sd0f98a098sda
Time-Out: Second-604800
Content-Type: text/xml
Content-Length: xxxxx

<?XML:Namespace href ="http://www.ietf.org/standards/dav/" AS = "D"/>
<D:Prop>
     <lockdiscovery>
          <activelock>
               <locktype>write</locktype>
               <lockscope>exclusive</lockscope>
               <addlocks/>
               <owner>
     <HREF>http://www.ics.uci.edu/~ejw/contact.html</HREF>
               </owner>
               <timeout>Second-604800</timeout>
               <locktoken>
                    <HREF>
               OpaqueLockToken:xyz122393481230912asdfa09s8df09s7d
               f08sd0f98a098sda
                    </HREF>
               </locktoken>
          </activelock>
     </lockdiscovery>
</D:Prop>

5.2.13 Example - Multi-Resource Lock Request

LOCK /workspace/webdav/proposal.doc HTTP/1.1
Host: webdav.sb.aol.com
Lock-Info: LockType=Write LockScope=Exclusive
Addlocks=<http://webdav.sb.aol.com/workspace/><http://foo.bar/blah>
Time-Out: Infinite, Second-4100000000
Owner: <http://www.ics.uci.edu/~ejw/contact.html>

HTTP/1.1 409 Conflict
Content-Type: text/xml
Content-Length: xxxxx

<?XML:Namespace href =
     "http://www.ietf.org/standards/dav/" As = "D"/>
<D:MultiResponse>
     <Response>
          <HREF>
          http://webdav.sb.aol.com/workspace/webdav/proposal.doc
          </HREF>
          <HREF>
          http://webdav.sb.aol.com/workspace/webdav/
          </HREF>
          <Status>HTTP/1.1 202 Accepted</Status>
     </Response>
     <Response>
          <HREF>http://foo.bar/blah</HREF>
          <Status>HTTP/1.1 403 Forbidden</Status>
     </Response>
</D:MultiResponse>

This lock request has failed, because the server rejected the lock request for http://foo.bar/blah. The 409 Conflict status code indicates that the server was unable to satisfy the request because there is a conflict between the state of the resources and the operation named in the request. Within the multiresponse, the 202 Accepted status code indicates that the lock method was accepted by the resources, and would have been completed if all resources named in the request were able to be locked. The 403 Forbidden status code indicates that the server does not allow lock requests on this resource.

5.3 Write Lock

5.3.1 Methods Restricted by Write Locks

Note, however, that as new methods are created it will be necessary to specify how they interact with a write lock.

5.3.2 Write Locks and Properties

If a property is write locked then a LOCK request on the associated resource MUST fail with a 409 "Conflict". Note that a write lock on a property MAY be extended to include the associated resource without the principal having explicitly requested the extension.

5.3.3 Write Locks and Null Resources

5.3.4 Write Locks and Collections

However, if a write lock request is issued to a collection containing internal member resources that are currently locked, the request MUST fail with a 409 Conflict status code.

5.3.5 Write Locks and COPY/MOVE

A COPY method invocation MUST NOT duplicate any write locks active on the source.

5.3.6 Re-issuing Write Locks

5.3.7 Write Locks and The State-Token Header

5.3.7.1 Problem Definition

5.3.7.2 Solution Requirement

The solution must not require that clients guess what sorts of locks might be used and use if-state-match headers with wildcards to prevent collisions. The problem with trying to "guess" which locks are being used is that new lock types might be introduced, and the program would not know to "guess them". So, for example, a client might put in an if-state-match header with a wildcard specifying that if any write lock is outstanding then the operation should fail. However a new read/write lock could be introduced which the client would not know to put in the header.

5.3.7.3 State-Token Header

If the following conditions are met: 1. a user-agent has authenticated itself as a principal, 2. the user-agent is submitting a method request to a resource on which the principal owns a write lock, 3. the method is restricted by a write lock, as defined in the section "Methods Restricted by a Write Lock", then the method request MUST include a State-Token header with the lock token of the write lock, or the method fails with a 409 Conflict status code. If multiple resources are involved with a method, such as a COPY or MOVE method, then the lock tokens, if any, for all involved resources, MUST be included in the State-Token request header.

For example, Program A, used by user A, takes out a write lock on a resource. Program A then makes a number of PUT requests on the locked resource, all the requests contain a State-Token header which includes the write lock state token. Program B, also run by User A, then proceeds to perform a PUT to the locked resource. However program B was not aware of the existence of the lock and so does not include the appropriate state-token header. The method is rejected even though principal A is authorized to perform the PUT. Program B can, if it so chooses, now perform lock discovery and obtain the lock token. Note that program A and B can perform GETs without using the state-token header because the ability to perform a GET is not affected by a write lock.

Having a lock state token provides no special access rights. Anyone can find out anyone else's lock state token by performing lock discovery. Locks are to be enforced based upon whatever authentication mechanism is used by the server, not based on the secrecy of the token values.

5.3.7.3.1 Example

COPY /~fielding/index.html HTTP/1.1
Host: www.ics.uci.edu
Destination: http://www.ics.uci.edu/users/f/fielding/index.html
State-Token: <OpaqueLockToken:123AbcEfg1284h23h2>
<OpaqueLockToken:AAAASDFcalkjfdas12312>

HTTP/1.1 200 OK

5.4 Lock Tokens

5.4.1 Problem Description

5.4.2 Lock Token Introduction

There are two types of lock tokens, a generic lock token, which is unique only for a particular resource, and an opaque lock token, which is unique across all resources for all time.

Uniqueness for a particular resource prevents problems with long held outstanding lock tokens being confused with newer tokens. This uniqueness requirement is the same as for e-tags. Uniqueness across all resources for all time allows for tokens to be submitted across resources and servers without fear of confusion.

Generic lock tokens, because of their relaxed uniqueness requirements, are faster to generate than opaque lock tokens.

5.4.3 Generic Lock Tokens

5.4.4 OpaqueLockToken Lock Token

All resources MUST recognize the opaquelocktoken scheme and be able to, at minimum, recognize that the lock token was not generated by the resource. Note, however, that resources are not required to generate opaquelocktokens.

In order to guarantee uniqueness across all resources for all time the opaquelocktoken requires the use of the GUID mechanism.

Opaquelocktoken generators however have a choice of how they create these tokens. They can either generate a new GUID for every lock token they create, which is potentially very expensive, or they can create a single GUID and then add extension characters. If the second method is selected then the program generating the extensions MUST guarantee that the same extension will never be used twice with the associated GUID.

Opaque-Lock-Token = "OpaqueLockToken" ":" GUID [Extension]
GUID = ; As defined in [LEACH]
Extension = *urlc   ;urlc is defined in [Berners-Lee et al.,
1997] (draft-fielding-url-syntax-07.txt)

5.5 UNLOCK Method

5.5.1 Problem Definition

5.5.2 Example

UNLOCK /workspace/webdav/info.doc HTTP/1.1
Host: webdav.sb.aol.com
State-Token: OpaqueLockToken:123AbcEfg1284h23h2

HTTP/1.1 200 OK

5.6 Discovery Mechanisms

5.6.1 Lock Capability Discovery

5.6.1.1 Problem Definition

5.6.1.2 SupportedLock Property

5.6.1.3 LOCKENTRY XML Element

5.6.1.4 LOCKTYPE XML Element

5.6.1.5 LOCKSCOPE XML Element

5.6.1.6 Example

PROPFIND  /container/ HTTP/1.1
Host: www.foo.bar
Content-Length: xxxx
Content-Type: text/xml

<?XML:Namespace href =
     "http://www.ietf.org/standards/dav/" AS = "D"/>
<D:PROPFIND>
     <prop><SupportedLock/></prop>
</D:PROPFIND>

HTTP/1.1 207 Multi-Response
Content-Type: text/xml
Content-Length: xxxxx

<?XML:Namespace href ="http://www.ietf.org/standards/dav/" AS = "D"/>
<D:MultiResponse>
     <Response>
          <Prop>
               <SupportedLock>
                    <LockEntry>
                         <LockType>Write</LockType>
                         <LockScope>Exclusive</LockScope>
                    </LockEntry>
                    <LockEntry>
                         <LockType>Write</LockType>
                         <LockScope>Shared</LockScope>
                    </LockEntry>
               </SupportedLock>
          </Prop>
          <Status>HTTP/1.1 200 Success</Status>
</Response>
</D:MultiResponse>

5.6.2 Active Lock Discovery

5.6.2.1 Problem Definition

5.6.2.2 Solution Requirements

5.6.2.3 LOCKDISCOVERY Property

Description: The LOCKDISCOVERY property returns a listing of who has a lock, what type of lock they have, the time out type and the time remaining on the time out, and the associated lock token. The server is free to withhold any or all of this information if the requesting principal does not have sufficient access rights to see the requested data. A server which supports locks MUST provide the LOCKDISCOVERY property on any resource with locks on it.

5.6.2.4 ACTIVELOCK XML Element

5.6.2.5 OWNER XML Element

5.6.2.6 TIMEOUT XML Element

5.6.2.7 ADDLOCKS XML Element

5.6.2.8 LOCKTOKEN XML Element

5.6.2.9 Example

PROPFIND  /container/ HTTP/1.1
Host: www.foo.bar
Content-Length: xxxx
Content-Type: text/xml

<?XML:Namespace href =
     "http://www.ietf.org/standards/dav/" AS = "D"/>
<D:PROPFIND>
     <prop><lockdiscovery/></prop>
</D:PROPFIND>

HTTP/1.1 207 Multi-Response
Content-Type: text/xml
Content-Length: xxxxx

<?XML:Namespace href ="http://www.ietf.org/standards/dav/" AS = "D"/>
<D:MultiResponse>
     <Response>
          <Prop>
               <lockdiscovery>
                    <activelock>
                         <locktype>write</locktype>
                         <lockscope>exclusive</lockscope>
                         <addlocks>
                              <HREF>http://foo.com/doc/</HREF>
                         </addlocks>
                         <owner>Jane Smith</owner>
                         <timeout>Infinite</timeout>
                         <locktoken>
                              <HREF>iamuri:unique!!!!!</HREF>
                         </locktoken>
                    </activelock>
               </lockdiscovery>
          </Prop>
          <Status>HTTP/1.1 200 Success</Status>
</Response>
</D:MultiResponse>

6 Version Control

7 Internationalization Support

8 Security Considerations

9 Copyright

This document and translations of it may be copied and furnished to others, and derivative works that comment on or otherwise explain it or assist in its implementation may be prepared, copied, published and distributed, in whole or in part, without restriction of any kind, provided that the above copyright notice and this paragraph are included on all such copies and derivative works. However, this document itself may not be modified in any way, such as by removing the copyright notice or references to the Internet Society or other Internet organizations, except as needed for the purpose of developing Internet standards in which case the procedures for copyrights defined in the Internet Standards process must be followed, or as required to translate it into languages other than English.

The limited permissions granted above are perpetual and will not be revoked by the Internet Society or its successors or assignees.

This document and the information contained herein is provided on an "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.

10 Acknowledgements

11 References

[Berners-Lee, 1997] T. Berners-Lee, "Metadata Architecture." Unpublished white paper, January 1997. http://www.w3.org/pub/WWW/DesignIssues/Metadata.html.

[Bradner, 1997] S. Bradner, "Key words for use in RFCs to Indicate Requirement Levels." RFC 2119, BCP 14. Harvard University. March, 1997.

[Bray, Sperberg-McQueen, 1997] T. Bray, C. M. Sperberg-McQueen, "Extensible Markup Language (XML): Part I. Syntax", WD-xml-lang.html, http://www.w3.org/pub/WWW/TR/WD-xml-lang.html.

[Connolly et al, 1997] D. Connolly, R. Khare, H.F. Nielsen, "PEP - an Extension Mechanism for HTTP", Internet draft, work-in-progress. draft-ietf-http-pep-04.txt, ftp://ds.internic.net/internet-drafts/draft-ietf-http-pep-04.txt.

[Fielding et al., 1997] R. Fielding, J. Gettys, J. Mogul, H. Frystyk, T. Berners-Lee, "Hypertext Transfer Protocol --HTTP/1.1." RFC 2068. U.C. Irvine, DEC, MIT/LCS. January, 1997. ftp://ds.internic.net/rfc/rfc2068.txt

[Lasher, Cohen, 1995] R. Lasher, D. Cohen, "A Format for Bibliographic Records," RFC 1807. Stanford, Myricom. June, 1995. ftp://ds.internic.net/rfc/rfc1807.txt

[Maloney, 1996] M. Maloney, "Hypertext Links in HTML." Internet draft (expired), work-in-progress, January, 1996.

[MARC, 1994] Network Development and MARC Standards, Office, ed. 1994. "USMARC Format for Bibliographic Data", 1994. Washington, DC: Cataloging Distribution Service, Library of Congress.

[Miller et al., 1996] J. Miller, T. Krauskopf, P. Resnick, W. Treese, "PICS Label Distribution Label Syntax and Communication Protocols" Version 1.1, W3C Recommendation REC-PICS-labels-961031. http://www.w3.org/pub/WWW/TR/REC-PICS-labels-961031.html.

[Slein et al., 1997] J. A. Slein, F. Vitali, E. J. Whitehead, Jr., D. Durand, "Requirements for Distributed Authoring and Versioning on the World Wide Web." Internet-draft, work-in-progress, draft-ietf-webdav-requirements-04.txt, ftp://ds.internic.net/internet-drafts/draft-ietf-webdav-requirements-04.txt.

[WebDAV, 1997] WEBDAV Design Team. "A Proposal for Web Metadata Operations." Unpublished manuscript. http://www.ics.uci.edu/~ejw/authoring/proposals/metadata.html

[Weibel et al., 1995] S. Weibel, J. Godby, E. Miller, R. Daniel, "OCLC/NCSA Metadata Workshop Report." http://purl.oclc.org/metadata/dublin_core_report.

[Yergeau, 1997] F. Yergeau, "UTF-8, a transformation format of Unicode and ISO 10646", Internet Draft, work-in-progress, draft-yergeau-utf8-rev-00.txt, http://www.internic.net/internet-drafts/draft-yergeau-utf8-rev-00.txt.

12 Authors' Addresses

Y. Y. Goland
Microsoft Corporation
One Microsoft Way
Redmond, WA 98052-6399
Email yarong@microsoft.com

E. J. Whitehead, Jr.
Dept. Of Information and Computer Science
University of California, Irvine
Irvine, CA 92697-3425
Email: ejw@ics.uci.edu

A. Faizi
Netscape
685 East Middlefield Road
Mountain View, CA 94043
Email: asad@netscape.com

S. R Carter
Novell
1555 N. Technology Way
M/S ORM F111
Orem, UT 84097-2399
Email srcarter@novell.com

D. Jensen
Novell
1555 N. Technology Way
M/S ORM F111
Orem, UT 84097-2399
Email dcjensen@novell.com