The Session Initiation Protocol (SIP) defines the SIP Uniform Resource Identifier (URI) as any resource to which a SIP request can be generated for the purposes of establishing some form of communications operation. These URIs can represent users (for example, sip:joe@example.com). The SIP URI can also represent a service, such as voicemail, conferencing, or a presence list. A common pattern across such SIP services is that the service is defined, and associated with a URI. In order to operate, that service needs to make use of a list of users (or, more generally, a list of resources). When a SIP request is sent to the service URI, the server providing the service reads that list, and then performs some kind of operation against each resource on the list. This is shown in .

| SIP | ---------------> | Service | --------> service | | URI | | --------> +-------------+ ]]> One important example of such a service is a presence list service. A presence list service allows a client to generate a SIP SUBSCRIBE request to ask for presence information for a list of users. The presence list server obtains the presence for the users on the list and provides them back to the client. A presence list server is a specific case of a resource list server (RLS) , which allows a client to generate a SIP SUBSCRIBE request to ask for notifications of SIP events for a list of resources. Another example of such a service is an instant conference service. If a client sends a SIP INVITE request to the URI representing the instance conference service, the conference server will create a conference call containing the client and the associated group of users. It is very useful for a user of these systems to define the groups of users or resources (generally called a resource list) separately from the services that access those resource lists. Indeed, there are usages for resource lists even in the absence of any associated network-based service. As an example, rather than use a presence list service, a client might generate individual SUBSCRIBE requests to obtain the presence of each user in a locally stored presence list. In such a case, there is a need for a format for storing the list locally on disk. Furthermore, the user might wish to share the list with friends, and desire to email it to those friends. This also requires a standardized format for the resource list. As such, this document defines two Extensible Markup Language (XML) document formats. The first is used to represent resource lists, independent of any particular service. The second is used to define service URIs for an RLS, and to associate a resource list with the service URI. This document also defines an XML Configuration Access Protocol (XCAP) application usage for managing each of these two documents.

In this document, the key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" are to be interpreted as described in RFC 2119 and indicate requirement levels for compliant implementations.

A resource lists document is an XML document that MUST be well-formed and MUST be valid according to schemas, including extension schemas, available to the validater and applicable to the XML document. Resource lists documents MUST be based on XML 1.0 and MUST be encoded using UTF-8. This specification makes use of XML namespaces for identifying resource lists documents and document fragments. The namespace URI for elements defined by this specification is a URN that uses the namespace identifier 'ietf' defined by RFC 2648 and extended by RFC 3688. This URN is: urn:ietf:params:xml:ns:resource-lists A resource lists document has the <resource-lists> element as the root element of the document. This element has no attributes. Its content is a sequence of zero or more <list> elements, each of which defines a single resource list. Each <list> element can contain an optional "name" attribute. This attribute is a handle for the list. When present, it MUST be unique amongst all other <list> elements within the same parent element. The <list> element may also contain attributes from other namespaces, for the purposes of extensibility. Each <list> element is composed of an optional display name, a sequence of zero or more elements, each of which may be an <entry> element, a <list> element, an <entry-ref> element, or an <external> element, followed by any number of elements from other namespaces, for the purposes of extensibility. The ability of a <list> element to contain other <list> elements means that a resource list can be hierarchically structured. The <display-name> then allows for a human-friendly name to be associated with each level in the hierarchy. An <entry> element describes a single resource, defined by a URI, that is part of the list. An <entry-ref> element allows an entry in a document within the same XCAP root to be included by reference, rather than by value. An <external> element contains a reference to a list stored on this or another server. The <entry> element describes a single resource. The <entry> element has a single mandatory attribute, "uri". This attribute is equal to the URI that is used to access the resource. The resource list format itself does not constrain the type of URI that can be used. However, the service making use of the resource list may require specific URI schemes. For example, RLS services will require URIs that represent subscribeable resources. This includes the SIP and pres URIs. The "uri" attribute MUST be unique amongst all other "uri" attributes in <entry> elements within the same parent. Uniqueness is determined by case-sensitive string comparisons. As such, it is possible that two "uri" attributes will have the same URI when compared using the functional equality rules defined for that URI scheme, but different ones when compared using case sensitive string comparison. The <entry> element can also contain attributes from other namespaces for the purposes of extensibility. The <entry> element contains a sequence of elements that provide information about the entry. Only one such element is defined at this time, which is <display-name>. This element provides a UTF-8-encoded string, meant for consumption by a human user, that describes the resource. Unlike the "name" attribute of the <entry> element, the <display-name> has no uniqueness requirements. The <display-name> element can contain the "xml:lang" attribute, which provides the language of the display name. The <entry> element can contain other elements from other namespaces. This is meant to support the inclusion of other information about the entry, such as a phone number or postal address. The <entry-ref> element allows an entry to be included in the list by reference, rather than by value. This element is only meaningful when the document was obtained through XCAP. In such a case, the referenced entry has to exist within the same XCAP root. The <entry> element has a single mandatory attribute, "ref". The "ref" attribute MUST be unique amongst all other "ref" attributes in <entry-ref> elements within the same parent. Uniqueness is determined by case sensitive string comparisons. The <entry-ref> element also allows attributes from other namespaces, for the purposes of extensibility. The content of an <entry-ref> element is an optional display name, followed by any number of elements from other namespaces, for the purposes of extensibility. The display name is useful for providing a localized nickname as an alternative to the name defined in the <entry> to which the <entry-ref> refers. The content of the "ref" attribute is a relative HTTP URI . Specifically, it MUST be a relative path reference, where the base URI is equal to the XCAP root URI of the document in which the <entry-ref> appears. This relative URI, if resolved into an absolute URI according to the procedures in RFC 3986, MUST resolve to an <entry> element within a resource-lists document. For example, suppose that an <entry> element within a specific XCAP root was identified by the following HTTP URI:

If http://xcap.example.com is the XCAP root URI, then an <entry-ref> element pointing to this entry would have the following form:

]]> Note that line folding within the HTTP URI and XML attribute above are for the purposes of readability only. Also note that, as described in RFC 3986, the relative path URI does not begin with the "/". Since the relative URI used within the "ref" attribute must be a relative path URI, the "/" will never be present as the first character within the content of a "ref" attribute. Since the content of the "ref" attribute is a valid HTTP URI, it must be percent-encoded within the XML document. The <external> element is similar to the <entry-ref> element. Like <entry-ref>, it is only meaningful in documents obtained from an XCAP server. It too is a reference to content stored elsewhere. However, it refers to an entire list, and furthermore, it allows that list to be present on another server. The <external> element has a single mandatory attribute, "anchor", which specifies the external list by means of an absolute HTTP URI. The "anchor" attribute MUST be unique amongst all other "anchor" attributes in <external> elements within the same parent. Uniqueness is determined by case-sensitive string comparisons. The <external> element can also contain attributes from other namespaces, for the purposes of extensibility. The content of an <external> element is an optional <display-name> followed by any number of elements from another namespace, for the purposes of extensibility. The value of the "anchor" attribute MUST be an absolute HTTP URI. This URI MUST identify an XCAP resource, and in particular, it MUST represent a <list> element within a resource lists document. The URI MUST be percent-encoded. For both the <entry-ref> and <external> elements, the responsibility of resolving their references falls upon the entity that is making use of the document. When the document is used in conjunction with XCAP, this means that the burden falls on the XCAP client. If the XCAP client is a PC-based application using the resource-lists document as a presence list, the references would likely be resolved upon explicit request by the user. They can, of course, be resolved at any time. If the XCAP client is an RLS itself, the references would be resolved when the RLS receives a SUBSCRIBE request for an RLS service associated with a resource list that contains one of these references (see below). An XCAP server defined by this specification will not attempt to resolve the references before returning the document to the client. Similarly, if, due to network errors or some other problem, the references cannot be resolved, the handling is specific to the usage of the document. For resource lists being used by RLS services, the handling is discussed below.

]]>

The following is an example of a document compliant to the schema. All line feeds within element content are for display purposes only.

Bill Doe

Close Friends

Joe Smith

Nancy Gross

Marketing

]]>

Resource lists documents can be manipulated with XCAP. This section provides the details necessary for such a usage.

XCAP requires application usages to define an application unique ID (AUID) in either the IETF tree or a vendor tree. This specification defines the "resource-lists" AUID within the IETF tree, via the IANA registration in .

The MIME type for this document is "application/resource-lists+xml".

The XML Schema for this document is defined as the sole content of .

The default namespace used in expanding URIs is urn:ietf:params:xml:ns:resource-lists.

In addition to the schema, there are constraints on the values present in the "name" attribute of the <list> element, the "uri" attribute of the <external> element, the "ref" attribute of the <entry-ref> element, and the "anchor" attribute of the <external> element. These constraints are defined in . Some of these constraints are enforced by the XCAP server. Those constraints are: The "name" attribute in a <list> element MUST be unique amongst all other "name" attributes of <list> elements within the same parent element. Uniqueness is determined by case-sensitive string comparison. The "uri" attribute in a <entry> element MUST be unique amongst all other "uri" attributes of <entry> elements within the same parent element. Uniqueness is determined by case-sensitive string comparison. The URI in the "ref" attribute of the <entry-ref> element MUST be unique amongst all other "ref" attributes of <entry-ref> elements within the same parent element. Uniqueness is determined by case-sensitive string comparison. The value of the attribute MUST be a relative path reference. Note that the server is not responsible for verifying that the reference resolves to an <entry> element in a document within the same XCAP root. The URI in the "anchor" attribute of the <external> element MUST be unique amongst all other "anchor" attributes of <external> elements within the same parent element. Uniqueness is determined by case-sensitive string comparison. The value of the attribute MUST be an absolute HTTP URI. Note that the server is not responsible for verifying that the URI resolves to a <list> element in a document. Indeed, since the URI may reference a server in another domain, referential integrity cannot be guaranteed without adding substantial complexity to the system.

Semantics for the document content are provided in .

Resource lists documents are usually identified as references from other application usages. For example, an RLS services document contains a reference to the resource list it uses. Frequently, an XCAP client will wish to insert or remove an <entry>, <entry-ref>, or <external> element from a document without having a cached copy of that document. In such a case, the "uri" attribute of the <entry> element, the "ref" attribute of the <entry-ref> element, or the "anchor" attribute of the <external> element is used as an index to select the element to operate upon. The XCAP server will determine uniqueness by case-sensitive string comparison. However, each of these attributes contain URIs, and the URI equality rules for their schemes may allow two URIs to be the same, even if they are different by case sensitive string comparison. As such, it is possible that a client will attempt a PUT or DELETE in an attempt to modify or remove an existing element. Instead, the PUT ends up inserting a new element, or the DELETE ends up returning an error response. If the XCAP client cannot determine whether the user intent is to create or replace, the client SHOULD canonicalize the URI before performing the operation. For a SIP URI (often present in the "uri" attribute of the <entry> element), this canonicalization procedure is defined in . We expect that the SIP URIs that will be placed into resource lists documents will usually be of the form sip:user@domain, and possibly include a user parameter. The canonicalization rules work perfectly for these URIs. For HTTP URIs, a basic canonicalization algorithm is as follows. If the port in the URI is equal to the default port (80 for http URIs), then the port is removed. The hostname is converted to all lowercase. Any percent-encoding in the URI for characters which do not need to be percent-encoded is removed. A character needs to be percent-encoded when it is not permitted in that part of the URI based on the grammar for that part of the URI.

There are no resource interdependencies identified by this application usage.

This application usage does not modify the default XCAP authorization policy, which is that only a user can read, write, or modify their own documents. A server can allow privileged users to modify documents that they don't own, but the establishment and indication of such policies is outside the scope of this document. It is anticipated that a future application usage will define which users are allowed to modify a list resource.

An RLS services document is used to define URIs that represent services provided by a Resource List Server (RLS) as defined in . An RLS services document is an XML document that MUST be well-formed and MUST be valid according to schemas, including extension schemas, available to the validater and applicable to the XML document. RLS services documents MUST be based on XML 1.0 and MUST be encoded using UTF-8. This specification makes use of XML namespaces for identifying RLS services documents and document fragments. The namespace URI for elements defined by this specification is a URN that uses the namespace identifier 'ietf' defined by RFC 2648 and extended by RFC 3688. This URN is: urn:ietf:params:xml:ns:rls-services The root element of an rls-services document is <rls-services>. It contains a sequence of <service> elements, each of which defines a service available at an RLS. Each <service> element has a single mandatory attribute, "uri". This URI defines the resource associated with the service. That is, if a client subscribes to that URI, they will obtain the service defined by the corresponding <service> element. The <service> element can also contain attributes from other namespaces, for the purposes of extensibility. The <service> element contains child elements that define the service. For an RLS service, very little service definition is needed: just the resource list to which the server will perform virtual subscriptions and the set of event packages that the service supports. The former can be conveyed in one of two ways. There can be a <resource-list> element, which points to a <list> element in a resource-lists document, or there can be a <list> element, which includes the resource list directly. The supported packages are contained in the <packages> element. The <service> element can also contain elements from other namespaces, for the purposes of extensibility. By including the contents of the resource list directly, a user can create lists and add members to them with a single XCAP operation. However, the resulting list becomes "hidden" within the RLS service definition, and is not usable by other application usages. For this reason, the <resource-list> element exists as an alternative. It can reference a <list> element in a resource-lists document. Since the list is separated from the service definition, it can be easily reused by other application usages. The <list> element is of the list type defined by the schema for resource lists. It is discussed in . The <resource-list> element contains a URI. This element is only meaningful when the document was obtained through XCAP. The URI MUST be an absolute HTTP URI representing an XCAP element resource. Its XCAP root MUST be the same as the XCAP root of the RLS services document. When the RLS services document is present in a user's home directory, the HTTP URI MUST exist underneath that user's home directory in the resource-lists application usage. When the RLS services document is in the global directory, the HTTP URI MUST exist underneath any user's home directory in the resource-lists application usage. In either case, the element referenced by the URI MUST be a <list> element within a resource-lists document. All of these constraints except for the latter one (which is a referential integrity constraint) will be enforced by the XCAP server. The <packages> element contains a sequence of <package> elements. The content of each <package> element is the name of a SIP event package . The <packages> element may also contain elements from additional namespaces, for the purposes of extensibility. The <packages> element is optional. When it is not present, it means that the RLS service will accept subscriptions for any event package.

]]>

This document shows two services. One is sip:mybuddies@example.com, and the other is sip:marketing@example.com. The former service references a resource list in a resource-lists document, and the latter one includes a list locally. Both services are for the presence event package only.

http://xcap.example.com/resource-lists/user s/sip:joe@example.com/index/~~/resource-lists/list%5b@nam e=%22l1%22%5d

presence presence ]]>

RLS services documents can be manipulated with XCAP. This section provides the details necessary for such a usage.

XCAP requires application usages to define an application unique ID ID (AUID) in either the IETF tree or a vendor tree. This specification defines the "rls-services" AUID within the IETF tree, via the IANA registration in .

The MIME type for this document is "application/rls-services+xml".

The XML Schema for this document is defined as the sole content of .

The default namespace used in expanding URIs is urn:ietf:params:xml:ns:rls-services.

In addition to the schema, there are constraints on the URIs present in the <service> and <resource-list> elements. These constraints are defined in . Some of these constraints are enforced by the XCAP server. Those constraints are: The URI in the "uri" attribute of the <service> element MUST be unique amongst all other URIs in "uri" elements in any <service> element in any document on a particular server. This uniqueness constraint spans across XCAP roots. Furthermore, the URI MUST NOT correspond to an existing resource within the domain of the URI. If a server is asked to set the URI to something that already exists, the server MUST reject the request with a 409, and use the mechanisms defined in to suggest alternate URIs that have not yet been allocated. The URI in a <resource-list> element MUST be an absolute URI. The server MUST verify that the URI path contains "resource-lists" in the path segment corresponding to the AUID. If the RLS services document is within the XCAP user tree (as opposed to the global tree), the server MUST verify that the XUI in the path is the same as the XUI in the URI of to the RLS services document. These checks are made by examining the URI value, as opposed to dereferencing the URI. The server is not responsible for verifying that the URI actually points to a <list> element within a valid resource lists document. In addition, an RLS services document can contain a <list> element, which in turn can contain <entry>, <entry-ref>, <list>, and <external> elements. The constraints defined for these elements in MUST be enforced. In some cases, an XCAP client will wish to create a new RLS service, and wish to assign it a "vanity URI", such as sip:friends@example.com. However, the client does not know whether this URI meets the uniqueness constraints defined above. In that case, it can simply attempt the creation operation, and if the result is a 409 that contains a detailed conflict report with the <uniqueness-failure> element, the client knows that the URI could not be assigned. It can then retry with a different vanity URI, or use one of the suggestions in the detailed conflict report. If the client wishes to create a new RLS service, and it doesn't care what the URI is, the client creates a random one, and attempts the creation operation. As discussed in , if this should fail with a uniqueness conflict, the client can retry with different URIs with increasing randomness.

Semantics for the document content are provided in .

Typically, there are two distinct XCAP clients that access RLS services documents. The first is a client acting on behalf of the end user in the system. This client edits and writes both resource lists and RLS services documents as they are created or modified by the end user. The other XCAP client is the RLS itself, which reads the RLS services documents in order to process SUBSCRIBE requests. To make it easier for an RLS to find the <service> element for a particular URI, the XCAP server maintains, within the global tree, a single RLS services document representing the union of all the <service> elements across all documents created by all users within the same XCAP root. There is a single instance of this document, and its name is "index". Thus, if the root services URI is http://xcap.example.com, the following is the URI that an RLS would use to fetch this index:

As discussed below, this index is created from all the documents in the user tree that have the name "index" as well. An implication of this is that a client operating on behalf of a user SHOULD define its RLS services within the document named "index". If the root services URI is http://xcap.example.com, for user "sip:joe@example.com" the URI for this document would be:

If a client elects to define RLS services in a different document, this document will not be "picked up" in the global index, and therefore, will not be used as an RLS service.

As with other application usages, the XML schema and the XCAP resource naming conventions describe most of the resource interdependencies applicable to this application usage. This application usage defines an additional resource interdependence between a single document in the global tree and all documents in the user tree with the name "index". This global document is formed as the union of all of the index documents for all users within the same XCAP root. In this case, the union operation implies that each <service> element in a user document will also be present as a <service> element in the global document. The inverse is true as well. Every <service> element in the global document exists within a user document within the same XCAP root. As an example, consider the RLS services document for user sip:joe@example.com:

http://xcap.example.com/resource-lists/users/si p:joe@example.com/index/~~/resource-lists/list%5b@name=%22l1% 22%5d

presence ]]> And consider the RLS services document for user bob:

presence ]]> The global document at http://xcap.example.com/rls-services/global/index would look like this:

http://xcap.example.com/resource-lists/user s/sip:joe@example.com/index/~~/resource-lists/list%5b@nam e=%22l1%22%5d

presence presence ]]> Requests made against the global document MUST generate responses that reflect the most recent state of all the relevant user documents. This requirement does not imply that the server must actually store this global document. It is anticipated that most systems will dynamically construct the responses to any particular request against the document resource. The uniqueness constraint on the "uri" attribute of <service> will ensure that no two <service> elements in the global document have the same value of that attribute.

This application usage does not modify the default XCAP authorization policy, which is that only a user can read, write, or modify thier own documents. A server can allow privileged users to modify documents that they don't own, but the establishment and indication of such policies are outside the scope of this document. It is anticipated that a future application usage will define which users are allowed to modify an RLS services document. The index document maintained in the global tree represents sensitive information, as it contains the union of all the information for all users on the server. As such, its access MUST be restricted to trusted elements within domain of the server. Typically, this would be limited to the RLSs that need access to this document.

This section discusses how an RLS, on receipt of a SUBSCRIBE request, uses XCAP and the RLS services document to guide its operation. When an RLS receives a SUBSCRIBE request for a URI (present in the Request URI), it obtains the <service> element whose uri attribute matches (based on URI equality) the URI in the SUBSCRIBE request. This document makes no normative statements on how this might be accomplished. The following paragraph provides one possible approach. The RLS canonicalizes the Request URI as described in . It then performs an XCAP GET operation against the URI formed by combining the XCAP root with the document selector of the global index with a node selector of the form "rls-services/service[@uri=<canonical-uri>]", where <canonical-uri> is the canonicalized version of the Request URI. If the response is a 200 OK, it will contain the service definition for that URI. Once the <service> element has been obtained, it is examined. If the <packages> element is present, and the event package in the SUBSCRIBE request is not amongst those listed in the <package> elements within <packages>, the request MUST be rejected with a 489 (Bad Event) response code, as described in . Otherwise, it SHOULD be processed. The next step is to authorize that the client is allowed to subscribe to the resource. This can be done using the data defined in , for example. Assuming the subscriber is authorized to subscribe to that resource, the subscription is processed according to the procedures defined in . This processing requires the RLS to compute a flat list of URIs that are to be subscribed to. If the <service> element had a <list> element, it is extracted. If the <service> element had a <resource-list> element, its URI content is dereferenced. The result should be a <list> element. If it is not, the request SHOULD be rejected with a 502 (Bad Gateway). Otherwise, that <list> element is extracted. At this point, the RLS has a <list> element in its possession. The next step is to obtain a flat list of URIs from this element. To do that, it traverses the tree of elements rooted in the <list> element. Before traversal begins, the RLS initializes two lists: the "flat list", which will contain the flat list of the URI after traversal, and the "traversed list", which contains a list of HTTP URIs in <external> elements that have already been visited. Both lists are initially empty. Next, tree traversal begins. A server can use any tree-traversal ordering it likes, such as depth-first search or breadth-first search. The processing at each element in the tree depends on the name of the element: If the element is <entry>, the URI in the "uri" attribute of the element is added to the flat list if it is not already present (based on case-sensitive string equality) in that list, and the URI scheme represents one that can be used to service subscriptions, such as SIP and pres . If the element is an <entry-ref>, the relative path reference making up the value of the "ref" attribute is resolved into an absolute URI. This is done using the procedures defined in Section 5.2 of RFC 3986, using the XCAP root of the RLS services document as the base URI. This absolute URI is resolved. If the result is not a 200 OK containing a <entry> element, the SUBSCRIBE request SHOULD be rejected with a 502 (Bad Gateway). Otherwise, the <entry> element returned is processed as described in the previous step. If the element is an <external> element, the absolute URI making up the value of the "anchor" attribute of the element is examined. If the URI is on the traversed list, the server MUST cease traversing the tree, and SHOULD reject the SUBSCRIBE request with a 502 (Bad Gateway). If the URI is not on the traversed list, the server adds the URI to the traversed list, and dereferences the URI. If the result is not a 200 OK containing a <list> element, the SUBSCRIBE request SHOULD be rejected with a 502 (Bad Gateway). Otherwise, the RLS replaces the <external> element in its local copy of the tree with the <list> element that was returned, and tree traversal continues. Because the <external> element is used to dynamically construct the tree, there is a possibility of recursive evaluation of references. The traversed list is used to prevent this from happening. Once the tree has been traversed, the RLS can create virtual subscriptions to each URI in the flat list, as defined in . In the processing steps outlined above, when an <entry-ref> or <external> element contains a reference that cannot be resolved, failing the request is at SHOULD strength. In some cases, an RLS may provide better service by creating virtual subscriptions to the URIs in the flat list that could be obtained, omitting those that could not. Only in those cases should the SHOULD recommendation be ignored.

This section provides a technique for URI canonicalization. This canonicalization produces a URI that, in most cases, is equal to the original URI (where equality is based on the URI comparison rules in RFC 3261). Furthermore, the canonicalized URI will usually be lexically equivalent to the canonicalized version of any other URI equal to the original. To canonicalize the URI, the following steps are followed: First, the domain part of the URI is converted into all lowercase, and any tokens (such as "user" or "transport" or "udp") are converted to all lowercase. Secondly, any percent-encoding in the URI for characters which do not need to be percent-encoded is removed. A character needs to be percent-encoded when it is not permitted in that part of the URI based on the grammar for that part of the URI. For example, if a SIP URI is sip:%6aoe%20smith@example.com, it is changed to sip:joe%20smith@example.com. In the original URI, the character 'j' was percent-encoded. This is allowed, but not required, since the grammar allows a 'j' to appear in the user part. As a result, it appears as 'j' after this step of canonicalization. Thirdly, any URI parameters are reordered so that they appear in lexical order based on parameter name. The ordering of a character is determined by the US-ASCII numerical value of that character, with smaller numbers coming first. Parameters are ordered with the leftmost character as most significant. For parameters that contain only letters, this is equivalent to an alphabetical ordering. Finally, any header parameters are discarded. This canonicalized URI is used instead of the original URI. If two URIs, A and B, are functionally equal (meaning that they are equal according to the URI comparison rules in RFC 3261), their canonicalized URIs are equal under case-sensitive string comparison if the following are true: Neither URI contains header parameters. If one of the URI contains a URI parameter not defined in RFC 3261, the other does as well.

Resource-lists and RLS services documents are meant to be extended. An extension takes place by defining a new set of elements in a new namespace, governed by a new schema. Every extension MUST have an appropriate XML namespace assigned to it. The XML namespace of the extension MUST be different from the namespaces defined in this specification. The extension MUST NOT change the syntax or semantics of the schemas defined in this document. All XML tags and attributes that are part of the extension MUST be appropriately qualified so as to place them within that namespace. This specification defines explicit places where new elements or attributes from an extension can be placed. These are explicitly indicated in the schemas by the <any> and <anyAttribute> elements. Extensions to this specification MUST specify where their elements can be placed within the document. As a result, a document that contains extensions will require multiple schemas in order to determine its validity: a schema defined in this document, along with those defined by extensions present in the document. Because extensions occur by adding new elements and attributes governed by new schemas, the schemas defined in this document are fixed and would only be changed by a revision to this specification. Such a revision, should it take place, would endeavor to allow documents compliant to the previous schema to remain compliant to the new one. As a result, the schemas defined here don't provide explicit schema versions, as this is not expected to be needed.

The information contained in rls-services and resource-lists documents are particularly sensitive. It represents the principle set of people with whom a user would like to communicate. As a result, clients SHOULD use TLS when contacting servers in order to fetch this information. Note that this does not represent a change in requirement strength from XCAP.

There are several IANA considerations associated with this specification.

This section registers two new XCAP Application Unique IDs (AUIDs) according to the IANA procedures defined in .

resource-lists A resource lists application is any application that needs access to a list of resources, identified by a URI, to which operations, such as subscriptions, can be applied.

rls-services A Resource List Server (RLS) services application is a Session Initiation Protocol (SIP) application whereby a server receives SIP SUBSCRIBE requests for resource, and generates subscriptions towards a resource list.

This specification requests the registration of two new MIME types according to the procedures of RFC 4288 and guidelines in RFC 3023.

application resource-lists+xml none Same as charset parameter application/xml as specified in RFC 3023. Same as encoding considerations of application/xml as specified in RFC 3023. See Section 10 of RFC 3023 and of RFC 4826. none RFC 4826 This document type has been used to support subscriptions to lists of users for SIP-based presence. Magic Number: none File Extension: .rl Macintosh file type code: "TEXT" Jonathan Rosenberg, jdrosen@jdrosen.net COMMON The IETF.

application rls-services+xml none Same as charset parameter application/xml as specified in RFC 3023. Same as encoding considerations of application/xml as specified in RFC 3023. See Section 10 of RFC 3023 and of RFC 4826. none RFC 4826 This document type has been used to support subscriptions to lists of users for SIP-based presence. Magic Number: none File Extension: .rs Macintosh file type code: "TEXT" Jonathan Rosenberg, jdrosen@jdrosen.net COMMON The IETF.

This section registers two new XML namespaces, as per the guidelines in RFC 3688.

The URI for this namespace is urn:ietf:params:xml:ns:resource-lists. IETF, SIMPLE working group, (simple@ietf.org), Jonathan Rosenberg (jdrosen@jdrosen.net).

Resource Lists Namespace

Namespace for Resource Lists

urn:ietf:params:xml:ns:resource-lists

See RFC4826.

END ]]>

The URI for this namespace is urn:ietf:params:xml:ns:rls-services. IETF, SIMPLE working group, (simple@ietf.org), Jonathan Rosenberg (jdrosen@jdrosen.net).

Resource List Server (RLS) Services Namespace

Namespace for Resource List Server (RLS) Services

urn:ietf:params:xml:ns:rls-services

See RFC4826.

END ]]>

This section registers two XML schemas per the procedures in .

urn:ietf:params:xml:schema:resource-lists IETF, SIMPLE working group, (simple@ietf.org), Jonathan Rosenberg (jdrosen@jdrosen.net). The XML for this schema can be found as the sole content of .

urn:ietf:params:xml:schema:rls-services IETF, SIMPLE working group, (simple@ietf.org), Jonathan Rosenberg (jdrosen@jdrosen.net). The XML for this schema can be found as the sole content of .

The authors would like to thank Hisham Khartabil, Jari Urpalainen, and Spencer Dawkins for their comments and input. Thanks to Ted Hardie for his encouragement and support of this work.