There are a number of MIX roles for each channel, listed in the following table. Rights will be assigned to the various roles in the channel configuration node.

There MUST always be at least one Owner for a Channel. Owners, Administrators, Participants, and Allowed are optional and do not need to be set. Where no owner is explicitly set, it is anticipated that a server administrator will have owner rights. Rights are defined in a strictly hierarchical manner following the order of this table, so that for example Owners will always have rights that Administrators have.

Nodes MAY be archived and where this is done MAM MUST be used. Archiving of the Messages node MUST be done as part of the MIX specification. Archiving of other nodes is OPTIONAL.

The Messages node is used to distribute messages. The Messages node is a transient node and so no PubSub items are held. Messages MUST go to the associated MAM archive and history is retrieved by use of MAM. Users subscribe to this node to indicate that messages from the channel are to be sent to them. Private Messages are not distributed by the Messages node.

Each channel participant is represented as an item of the 'urn:xmpp:mix:nodes:participants' channel node. Each item is named by the bare proxy JID of the participant. For example '123456#coven@mix.shakespeare.example' might name the node item associated with participant 'hag66@shakespeare.example'. Information is stored in a <participant/> element qualified by the 'urn:xmpp:mix:1' namespace. The nick associated with the user is optional and is stored in a <nick/> child element of the <participant/> element. The nick for each channel participant MUST be different to the nick of other participants.

When a user joins a channel, the user's bare proxy JID is added to the participants node by the MIX service. When a user leaves a channel, the user's bare proxy JID is removed from the participants node. The participants node MUST NOT be directly modified using pubsub.

Users MAY subscribe to and read information this node, when permitted by the channel. Standard PubSub access will allow a full list of participants and associated nicks to be determined. By subscribing to the node, a user will be informed of changes to the Participants Node.

The participants node is OPTIONAL. If the Participants Node is not used, information on channel participants is not shared. If there is no participants node, the access control value 'participants' MUST NOT be used. The Participants node is a permanent node with one item per participant.

The JID Map node is used to associate a proxy bare JID to its corresponding real bare JID. It is a PubSub node with the 'node' attribute set to 'urn:xmpp:mix:nodes:jidmap'. The JID Map node MUST have one entry for each entry in the Participants node. This value is added when a user joins the channel and is removed when the user leaves the channel. Each item is identified by proxy bare JID, mapping to the real bare JID. This node is used to give administrator access to real JIDs and participant access to real JIDs in jid-visible channels. This node MUST NOT be modified directly using pubsub. In JID Visible channels, all participants MAY subscribe to this node. In JID Hidden and JID Maybe Visible channels, only administrators can subscribe. The JID Map node is a permanent node with one item per participant. Information is stored in a <participant/> element qualified by the 'urn:xmpp:mix:1' namespace. The real JID is stored in a <jid/> child element of the <participant/> element.

The JID Maybe Visible Map node is a similar node to the JID Map node that is used in addition to the JID Map Node in JID Maybe Visible channels. It is a PubSub node with the 'node' attribute set to 'urn:xmpp:mix:nodes:jidmap-visible'. All participants may subscribe to and access this node. It uses the same encoding as JID Map node and all participant JIDs MUST be included. Where a participant's preference is to not share the JID, the encoded participant value is the proxy JID. This will enable a user looking up a JID to clearly determine that the user preference is to not share the JID and to clearly distinguish this case from an erroneous proxy JID.

The presence node contains the presence value for clients belonging to participants that choose to publish presence to the channel. A MIX channel MAY require that all participants publish presence. Each item in the presence node is identified by the full proxy JID, and contains the current presence value for that JID. The presence is encoded in the same way as data that would be sent in a presence stanza using a <presence/> element qualified by the 'jabber:client' namespace. The full proxy JID is always used in this node. In MIX it is possible to have a 'presence-less channel' by not using this node. Access Control MAY be set to enforce that for each of the full JIDs in this list, the bare JID MUST be in the participants list.

This node MAY be subscribed to by users and this subscription MUST use the user's bare JID. So presence of online clients is sent to the user's server for each user subscribed to this node. Presence is always sent using standard presence protocol and NOT using pubsub protocol. Clients MUST NOT directly access the presence node using pubsub. The Presence node is a permanent PubSub node.

The Information node holds information about the channel. The information node contains a single item with the current information. The information node is named by the date/time at which the item was created. The information node is accessed and managed using standard pubsub. The Information node is a permanent node with a maximum of one item. Users MAY subscribe to this node to receive information updates. The Information node item MAY contain the following attributes, each of which is OPTIONAL:

Name and Description of the channel apply to the whole channel and will usually be changed infrequently.

JID visibility is included in the Information Node as this is information that will be useful for participant clients and may also be important when choosing to join a channel.

The name and description values MUST contain a "text" element and MAY contain additional text elements. Where multiple text elements are provided, each MUST posses an xml:lang attribute that describes the natural language of the element. The format of the Information node follows &xep0004;. This allows configuration to be updated by MIX defined commands and that the results of update commands are the same as the PubSub node. The following example shows the format of a item in the information node for the example coven@mix.shakespeare.example channel.

This node represents a list of JIDs that are allowed to become participants. If the Allowed node is not present, all JIDs are allowed. This node is accessed and managed using standard pubsub. The Allowed list is always considered in conjunction with the banned list, stored in the banned node. Only Administrators and Owners have write permission to the Allowed node and are also the only roles that are allowed to subscribe to this node. The Allowed node is a permanent node. Each item contains a real bare JID. The following example shows how the Allowed list can specify single JIDs and domains.

This node represents a list of JIDs that are explicitly not allowed to become participants. The values in this list take priority over values in the Allowed node. This node is accessed and managed using standard pubsub Only Administrators and Owners have write permission to the Banned node and are also the only roles that are allowed to subscribe to this node. Each item contains a real bare JID. The Banned node can contain bare JIDs and/or domains. The Banned node is a permanent node.

The Configuration node holds the configuration of the channel as a single item, named by the date-time of the last update to the configuration. The Configuration node is a permanent node with a maximum of one item. Previous configuration history MAY be accessed by MAM. Users with read access to the configuration node MAY subscribe to the configuration node to get notification of configuration change. This node is accessed and managed using standard pubsub. The configuration node is OPTIONAL for a MIX channel. For example, configuration choices could be fixed and not exposed. A subset of the defined configuration options MAY be used and additional non-standard configuration options MAY be added. JIDs in the configuration MUST be real bare JIDs and not proxy JIDs. If configuration options to control functionality of the nature described here are provided, the options defined in this standard MUST be used. The following configuration attributes are defined:

The configuration node is in &xep0004; format and includes all of the options used by the channel, including values for options using default values. This means that the value in the form can be directly mapped with the form returned by configuration administration commands. Configuration nodes will typically have a large number of elements. The following short example is provided to illustrate the syntax of the configuration node.

A user joins a channel by sending a MIX "join" command from one of the user's clients. There is no default set of nodes, so the user MUST specify the set of nodes to be subscribed to. To achieve the equivalent service to MUC, a user would subscribe to both messages and presence nodes. A user will typically subscribe to at least the message and/or presence nodes but MAY join and not subscribe to any nodes. The <join/> is a child element of <iq/> element. The <join/> element is qualified by the 'urn:xmpp:mix:1' namespace. The channel is specified by a 'channel' attribute in the <join/> element. The requested nodes are encoded as <subscribe/> child elements of the <join/> element. The join leads to the server subscribing the user to each of the requested nodes associated with the channel. The MIX service will also add the user to the participant list by injecting a new item into the "urn:xmpp:mix:nodes:participants" node automatically.

The default MIX model is that only channel participants are allowed to subscribe to nodes. A MIX channel MAY allow non-participant subscription. This will be handled by clients directly subscribing to the desired PubSub nodes.

The user's server needs to make roster changes as part of the join functionality. Because of this, the join and leave operations need to operate indirectly. For this reason the initial join request is sent to the local server with the MIX channel specified as an attribute to the join. The join is sent from a client identified by a full JID to the user's bare JID.

The user’s server will then pass the request to the MIX channel, with the request coming from the user's bare JID.

The channel responds to the user's sever with an IQ-result addressed to the user's bare JID. This stanza includes the nodes to which the user has been successfully subscribed, as well as the bare JID that will be used for the user in this channel and added to the participant list. The JID returned in the join is the user's bare proxy JID. The following example shows the result of the above request when the request is completed in full.

The user's server will then make roster modifications as set out in the next section. After making these changes, the user's server will send the join response back to the client that made the join request, identified by the full JID. This is illustrated below:

This response informs the client that initiated the join request that the request has been completed. Note that the roster changes described in the next section will lead to roster update information being sent to all of the user's online clients, so that all of the user's clients will be updated with the new MIX channel information.

If a user cannot be subscribed to one or more of the requested nodes (e.g., because the node does not exist), but can be subscribed to some the response simply lists the nodes successfully subscribed. If at least one node is requested and none of the nodes requested are successfully subscribed to, an error response is sent indicating the reason that the first node requested was not subscribed to. This error response will also include other nodes requested where subscription failed for the same reason.

The following response example shows a successful response to the initial request example where the participant is not subscribed to all nodes associated with the channel (in this case only messages, participants, and information). This example shows the message from the MIX channel to the user's server, which will be modified and sent to the client.

The channel also adds the user to the participants node and sends a notification to subscribers of the participants node. The following example shows such a notification.

The user that has been added to the channel is identified by the item id of the item added to the Participants node, which is the proxy JID of the new channel participant. Note that the <participant> element MUST NOT include a nick of the user being added. The nick MAY be set after the join.

At the same time the participant MUST be added to the JID Map node, to map from proxy JID to real JID. For a JID Maybe Visible channel, the participant MUST be added to the JID Maybe Visible Map node. The value in this node MUST reflect the user's visibility preference for the channel and MUST be updated to reflect any changes to this preference.

A user MAY subsequently modify subscription to nodes in a channel by sending a subscription modification request encoded as a <update-subscription/$gt; child element of <iq/> element. The <update-subscription/$gt; element is qualified by the 'urn:xmpp:mix:1' namespace. The requested notes are encoded as <subscribe/> child elements of the <update-subscription/$gt; element with the node name encoded as a 'node' attribute. This modification goes directly from client to MIX channel, as this change does not impact the roster and so does not need any local action. The following example shows subscription modification.

As part of the channel joining process, the user's server MUST add the MIX channel to the user's roster. This is done to share the user's presence status with the channel and channel participants subscribing to presence, when the user wishes this presence to be shared. These roster entries also enables the user's client to quickly determine which channels the user has joined. This roster entry will lead to the user's server correctly sending user's presence from all the user's MIX clients to the MIX channel. Where the user wishes to share presence, the roster subscription is configured with one way presence, as presence is sent to the MIX channel but no presence information is sent about the MIX channel to the user.

If the user does not wish to publish presence information to the channel, the user's server will add the roster entry with mode subscription=none. The roster entry will be present to record that the user has joined the channel, but it will not send presence information to the channel. The user's server MUST do this when the user has chosen Presence preference of 'not share' as described below. If the user changes the value of the preference, the server MUST modify subscription mode to reflect this.

The user's server MUST remove this roster entry when the user leaves the channel.

A channel MAY publish an Avatar following &xep0084;. A client MAY make use of this information to associate an Avatar with the roster entry for a channel.

A channel MAY store user preferences and parameters with each user. A user JID visibility preference has the following values:

1. 'default'. In this setting, the channel default value will be used. This value is used if another value is not explicitly set. This means JID is visible for a JID Visible channel and JID is hidden for JID Hidden and JID Maybe Visible channels.
2. 'never'. If this is set, JID will never be shared and user will be removed from the channel if its mode changes to JID Visible.
3. 'always'. If this is set, JID will always be shared and users will be removed from the channel if its mode changes to JID Hidden.
4. 'prefer not'. If this is set, JID will only be shared if mode is JID Visible.

The user MAY specify that no Private Messages are to be sent from the channel, and allow vCard retrieval.

The user MAY specify that presence is not to be shared, which will prevent the MIX Channel from sending a presence probe for the user on channel start-up. The user will also choose to not include the MIX channel in the user's roster, so that presence will not be updated. Where the channel configuration sets 'Participants Must Provide Presence', this variable MUST be set to 'Share'.

The following table sets out the standardized user preference values. A MIX implementation MAY use other values.

When joining a channel, the client MAY specify one or more preference options as a &xep0004; form. In the response, the server MUST specify all of the user preferences supported by the server, with default values if the user has not requested a different value. The following example shows joining a channel and setting a preference. The following example shows the message sent from the user's server to the MIX channel, which will have been preceded by a message from the user's client to the user's server.

The following example shows the result of a successful join, which also reports all the user preferences. This example shows the message coming from the MIX channel to the user's server, which will be sent on to the user.

The client MAY also query the channel in order to find out which user preferences are supported and the options available. This will allow users to set options not specified in the standard, by providing a form template in the result. The request is encoded as a <user-preference/> child element of <iq/>. <user-preference/> is qualified by the 'urn:xmpp:mix:1' namespace. The result is encoded as a form child element in the <user-preference/> element.

A user MAY modify preferences using the corresponding set operation. The set MAY specify values for some or all attributes. All attributes are returned in the result.

Users generally remain in a channel for an extended period of time. In particular the user remains as a participant the channel when the user goes offline. Note that this is different to &xep0045;, where the client leaves a room when going offline. So, leaving a MIX channel is a permanent action for a user across all clients. In order to leave a channel, a user sends a MIX "leave" command to the channel. The leave command is encoded as a <leave/> child element of <iq/> element. The <leave/> element is qualified by the 'urn:xmpp:mix:1' namespace, with the channel specified as a 'channel" attribute. When a user leaves the channel, the user's server will remove the channel from the user's roster. Leave commands are sent indirectly through the user's server, to enable roster removal. Leaving is initiated by a client request, as shown in the following example.

The user's server will then generate a matching request to the MIX channel.

The MIX channel will then remove the user from the channel, as described below. A response is sent to the user's server.

After receiving the confirmation that the user has left the MIX channel, the user's server will remove the MIX channel entry from the user's roster. The user's server will then notify the client that requested to leave.

When the user leaves the channel, the MIX service is responsible for unsubscribing the user from all nodes in the channel and for removing the user from the participants and presence list. If the user has online presence when the user leaves the channel, the change of presence status caused by removing the user's entry or entries from the presence node will ensure that subscribers to the presence node are correctly updated on presence status. Deletion from the participants and presence functions as if the item (channel participant) had been deleted using the PubSub retract mechanism with notification set to true. Notification of the deletion is sent to clients subscribed to the participants PubSub nodes, as shown in the example below.

Each participant of a channel MAY have a nick, which is how other users in the channel will see the user. In some cases a nick is not needed, for example where a user reads messages in a channel but does not send messages or share presence information. A nick MUST be present for a user to send a message to a channel or for a user's presence to be shared on a channel. There are four ways that a user's nick can be obtained. The choice of mechanism or mechanisms is dependent on channel policy:

1. The nick is registered with the user account in some way, for example as part of user provisioning with nick configured as an attribute in a directory service. For example, this could be chosen by corporate services that wish to ensure consistent nick values for a set of users and channels.
2. The nick is registered with the MIX service, as described in Registering a Nick .
3. The user explicitly sets the nick, as described in this section.
4. The MIX service generates the nick. In this case it is RECOMMENDED that the assigned nick is a UUID following &rfc4122;.

A user will typically set a nick when joining a channel and MAY update this nick from time to time. The user does this by sending a command to the channel to set the nick. This command is a <setnick/> child element of <iq/> element. The <setnick/> element is qualified by the 'urn:xmpp:mix:1' namespace. The nick is encoded as a <nick/> child element of the <setnick/> element. If the user wishes the channel to assign a nick (or knows that the channel will assign a nick) the nick field can be left blank, so that the user can see what is assigned in the result.

On successful nick assignment, the channel will return the nick that is to be used, noting that this MAY be different to the requested nick. MIX services SHOULD apply the "nickname" profile of the PRECIS OpaqueString class, as defined in &rfc7700;. The channel MAY return a conflict error or other appropriate error.

A nick MAY be associated with the user's bare JID. A user can register a nick with the MIX service. Nick registration can be used ensure that a user is able to use the same nick in all channels in the service and to prevent other users from using a registered nick. This can help achieve a consistent experience across a set of channels and prevent user confusion. Support for nick registration by a MIX service is OPTIONAL. Where nick registration is supported, nick registration MAY be OPTIONAL or MANDATORY. Where a user has registered a Nick with the MIX service, it MAY be used by each channel according to policy for the channel. A channel MAY enforce use of a registered nick. A channel MUST NOT use a registered nick for any other participant.

In order to determine if a Nick is allowed to be registered, the user MAY use disco to determine capabilities of the MIX service.

The response will be a list of features of the MIX channel. If Nick Registration is supported, then the result set will include <feature var="urn:xmpp:mix:1#nick-register"/>.

To register a nick with the MIX service the user sends a register command to the service. This is encoded as a <register/> child element of an <iq/> element. The <register/> element is qualified by the urn:xmpp:mix:1' namespace. The nick is encoded in a <nick/> child element of the <register/> element.

On success, the service informs the user of its nick. MIX SHOULD apply the "nickname" profile of the PRECIS OpaqueString class, as defined in &rfc7700; to the requested nick. This means that nick that is issued might be different from the nick that was requested.

When an Nick is assigned, the MIX server MUST update the Participants Node in the channel to reflect this change. Any users subscribed to this node will be notified of the change of Nick.

The following example shows an example of reporting successful Nick assignment.

If the requested nick is already taken and the MIX service does not assign an alternate nick, the MIX service MUST return a <conflict/> error:

If the register request does not contain a <nick/> element, then the MIX service MUST assign one. It is RECOMMENDED that the assigned nick is a UUID following &rfc4122;.

A user joins a channel over an extended period, and participation in a channel does not generally change when user goes online or offline. The user's participation in a channel is reflected by the user's bare JID in the participant node. All messages to the channel are sent to this JID.

A user MAY share presence information with the channel, for one or more online clients. Where a user shares presence information with a channel, the channel is entered by the user's server into the user's roster when the user subscribes to the channel. When an XMPP client comes online or changes presence status, this will be communicated by the user to the user's server using broadcast presence. The user's XMPP server is then responsible to share this presence information to each entry in the roster and in particular to each MIX channel in the roster. The MIX channel will update the "urn:xmpp:mix:nodes:presence" node with any change of status and the updated presence information and then share this updated presence with users subscribed to this node, as described below. When the user sets an explicit status, this is used to modify the presence node in the channel. When a client being used by the user goes offline, the associated server will send presence status "unavailable" to the MIX channel, which will cause the full JID for that client to be removed from the presence node.

A channel MAY require that all channel participants share presence information with the channel, which is represented in the "urn:xmpp:mix:nodes:presence" node. If sharing presences is needed by the channel, an XMPP client conforming to this specification MUST share presence with the channel by including the channel in the user's roster. Note that the MIX service cannot generally enforce this, but it can require and enforce that when a message is sent to the channel, that the sender of the message is in the presence list.

Presence status and availability is set in a MIX channel by standard presence stanzas sent to the MIX channel by the user's server. Users wishing to receive presence updates will subscribe to the MIX channel presence node. Presence updates are sent out to subscribing participants using standard presence stanzas.

A user setting status is now used as an example. Unlike in &xep0045; where coming online is a special action, coming online in MIX is implicit when presence status is set. Going offline is a achieved by setting presence status to unavailable, which removes the client full JID entry from the presence node. When a user sets a presence status, the user's server sends updated presence to the MIX channel, and the MIX service then publishes the user's availability to the "urn:xmpp:mix:nodes:presence" node. If there is not an item named by the full JID of the client with updated presence status, this item is created. The sequence is shown in the following examples, starting with a client setting presences status on the connected server.

The server then sends the presence information to roster entries. The following example then shows the presence message from the client's server to the MIX channel.

The user's presence information is then published by the service to the "urn:xmpp:mix:nodes:presence" node, with the 'publisher' attribute set to the user's participant identifier (the proxy JID). The MIX channel then broadcasts the presence change to all users who are subscribed to the "urn:xmpp:mix:nodes:presence" node. The presence stanza is sent from the full proxy JID of the client updating status. Note that presence is associated with a client and so will have a full JID. The following example shows a presence message as distributed by the server to a presences subscriber.

The presence is distributed to those subscribing to the MIX channel presence node using a standard XMPP presence stanza. The presence change is recorded on the "urn:xmpp:mix:nodes:presence" node. The history of this node will be held as PubSub format in the MAM archive, so that presence history can be viewed.

The presence information for a channel is stored in the urn:xmpp:mix:nodes:presence node and distributed using standard presence stanzas to the server of each user subscribing to this presence node. The user's local server will then pass this presence information on to all online clients. This ensures that an online client is kept updated with presence. When a client goes offline, it will cease getting presence updates. Presence updates will continue to flow to the user's local server, and so the local server is able maintain up to date presence state for the channel. The user's server MAY cache this presence information to optimize performance or MAY discard it.

When the client comes online, it will activate use of the MIX. The user's server will then send full presence status to the client using standard presence messages. This will enable the client to update presence information for the channel. Note that this does not need any interaction with the channel.

In normal operation a MIX participant's server will hold accurate presence status for the channel which it provides to clients when they are activated. Incoming presence updates are immediately sent to active MIX clients.

There are two situations where a MIX participant's server will need to get presence status from the channel. The first time is when a user joins the channel as a participant and subscribes to presence. Upon this subscription the MIX channel will send to the participant's server (using the user's bare JID) presence for all of the items in the presence node using standard presence stanzas. This will give the participant's full current presence for the channel.

The second scenario is when the MIX participant's server needs to load or refresh presence status for a channel. This will be needed when the participant's server is started. This MIX participant's server requests presence update by sending a directed presence stanza to the MIX channel from the user's bare JID. The MIX channel can distinguish this from a presence update, which will always be sent from the clients full JID. This special presence stanza will send to the participant's server (using the user's bare JID) presence for all of the items in the presence node using standard presence stanzas.

Presence information will provide a MIX client with the nicks and proxy JIDs for participants in a channel. Messages sent to a channel and retrieved from MAM archives will show the proxy JID and nick of the sender. It is sometimes useful to determine the real JID associated with proxy JID. This can always be done for JID Visible channels and can sometimes be done for JID Maybe Visible channels.

For current users JID visible rooms, the real JID is found by a PubSub lookup on the JID Map node. This is shown in the following example:

For JID Maybe Visible rooms the lookup is performed on the JID Maybe Visible Map node. Note that where a user prefers to not share real JID, the result of this lookup of proxy JID will be the (same) proxy JID.

When an older message is considered, it is possible that the proxy JID of the sender is not current. Such a JID can be looked up in the MAM Archive of the JID Map Node. This is shown in the following example:

A MIX client will typically display message history of the channel to the user. When a client comes online it will need to obtain this message history from the MAM archive associated with the channel. There are three basic approaches a client will take:

1. If the client has previously displayed message history and has been offline for a reasonably small time, the client MAY wish to retrieve all messages since the last one displayed to the user.
2. The client MAY wish to display a fixed number of messages, perhaps finding more messages if the user subsequently requests.
3. The client MAY wish to display messages from a recent time period, perhaps finding more messages if the user subsequently requests.

To achieve this, the client will query the user's own MAM archive using &xep0313;, with the query filtered by the channel JID. This gives the client flexibility to retrieve and display message history in a manner appropriate to the client implementation.

The only exception to this is when a user wishes to access message history in the channel prior to when the user joined the channel. To achieve this, the client will use MAM to retrieve message history directly from the MAM Archive of the MUX channel.

When a client goes offline, this presence update is sent by the client's server to the MIX channel. From the client perspective, this is the same as any other presence change. Handling by the MIX channel is slightly different.

The MIX channel will retract (remove) the item in the presence node of the MIX channel identified by the client's full JID. The MIX channel will notify subscribers to the presence node of the user going offline by sending a presence stanza to the full JID of each client. The presence stanza will reference the full proxy JID of the client that is going offline, as shown in the following example:

There is the possibility that the message associated with the user going offline will be lost. If this happens, "ghost" entries will appear in the presence node. A MIX service MAY take steps to address this, for example by probing client with a disco for presence items that remain unchanged for a long period.

It is desirable to prevent clients from going offline briefly and then coming back online again, as this will lead to "flapping presence". The RECOMMENDED approach to achieve this is use of &xep0198; to maintain an XMPP client connection in the event of short term TCP failure.

A client sends a message directly to a MIX channel as a standard groupchat message, in exactly the same way as for &xep0045;. Messages are sent directly to the MIX channel from the user's client. The message id is selected by the client.

The MIX channel then adds information to the message using a <mix> element qualified by the 'urn:xmpp:mix:1' namespace. This element contains two child elements:

1. A <nick> element that contains the Nick of the message sender, taken from the Participants Node.
2. A <jid> element containing the full proxy JID of the sender.

The MIX channel then puts a copy of the message into the MAM archive for the channel and sends a copy of the message to each participant in standard groupchat format. These messages sent by the channel are addressed to the bare JID of each participant and this will be handled by the participant's local server. The message 'from' attribute is the JID of the channel. The id of the message is the ID from the MAM archive and NOT the id used by the sender. The message placed in the MAM archive is the reflected message without a 'to' attribute.

The messages sent to participants have a different message id to the originally submitted message. This does not impact most recipients, but it does not allow the message originator to correlate the message with the submitted message. To address this the MIX channel MUST include an additional <submission-id> element in the <mix> element of the message copy going back to the originator's bare JID. The <submission-id> element holds the original id provided by the sender. This enables the originator to correlate the received message with the message submitted.

A MIX channel MAY support message retraction, where the sender of a messages or an authorized administrator deletes a message. A MIX channel MAY limit the time frame in which a message is allowed to be retracted, for example to prevent retraction of very old messages. When a messages is retracted the original message MAY be replaced by a tombstone. Message retraction is done by sending a special message that identifies the original message. This mechanism allows the retraction to be distributed on the same path as the original message so that all participating servers and clients MAY honour the retraction. The protocol to request retraction does this by adding to a message a <retract> element qualified by the 'urn:xmpp:mix:1' namespace. A retract messages MUST NOT have a body. The <retract> element MUST include an <id> attribute that holds the MAM-ID of the original message. The message sender will need to look up the MAM-ID. The MAM-ID is the convenient message identification for message recipients. A message and it's retraction shown in the following example.

The MIX channel will allow a user to retract a message sent by the user if the 'Allow User Message Retraction' option is configured. The MIX channel will allow an administrative user to retract any message if the user is in the group specified by the 'Administrator Message Retraction Rights' option.

If the retraction message is accepted, it MUST be distributed to channel participants. This will allow retraction to happen in the MAM archive of each channel participant and to reflect the retraction in client GUI. A client receiving a retraction message SHOULD ensure that the retracted message is no longer displayed to the end user.

Two approaches to message retraction can be used. In the first approach, the retracted message is simply removed. This is appropriate where retraction is provided as a user service and the user has rights to remove messages sent from the record.

The second approach is to leave a tombstone, which if taken MUST be done in the following manner. It is recommended to use a tombstone, as simply deleting the message may cause confusion with MAM queries. Use of a tombstone is appropriate where it is desired to leave a record of the message that was redacted. With this approach, the original message <body> is removed and replaced with a tombstone using the <retracted> element qualified by the 'urn:xmpp:mix:1' namespace that shows the JID of user performing the retraction and the time of the retraction.

A convenient way to reference another channel is to use &xep0372; which enables the JID of a channel to be shared. This might be used simply to inform the message recipient about the channel or as mechanism to invite the user to join the channel. This is useful as an invitation mechanism to a channel that any user can join or where the invitee knows that the user is allowed to join (e.g., because the channel is for all users in an organization).

Invitation by reference, as described in the previous section, is a convenient approach to invite a user to join a channel that the user has permission to join. This section describes the approach used when the inviter has permission to grant rights for the invitee to become a channel participant. This might be because the inviter is an administrator of the channel or the channel or if the inviter is a channel participant and the channel allows invitation by participants (this channel capability is controlled by the channel configuration variable 'Participation Addition by Invitation from Participant'). Invitation by reference is used to avoid cluttering the allowed node with JIDs of users who are invited to join, but do not accept the invitation. When a channel participant(the inviter) invites another user (the invitee) to join a channel, the following sequence of steps is followed:

1. The inviter checks with disco that the invitee supports MIX.
2. The channel inviter sends to the channel requesting an invite for the invitee.
3. The channel sends an invitation to the inviter.
4. The inviter sends the invitation to the invitee.
5. The invitee MAY use the invitation to join the channel.
6. The invitee MAY send a response to the inviter, indicating if the invitation was accepted or declined.

The first step is for the inviter to request an invitation from the channel. The invitation contains inviter, invitee and a token. The channel will evaluate if the inviter has rights to issue the invitation. This will be because the inviter is a channel administrator or if the inviter is a channel participant and the channel allows invitation by participants. If the inviter has rights to make the invitation, the channel will return a token. The token is a string that the channel can subsequently use to validate an invitation. The format of the token is not specified in this standard. The encoded token MAY reflect a validity time. The invitation request is encoded as an <invite/> child element of an <iq/> element. The <invite/> element is qualified by the 'urn:xmpp:mix:1' namespace. <invite/> contains an <invitation/> child element, which contain <inviter/>, <invitee/>, <channel/> and <token/> child elements.

The inviter can now send the invitee a message containing the invitation within the <message/> element, as shown in the following example.

The invitation can now be used by the invitee to join a channel. The <invitation/> child element is simply added to the standard channel <join/> element, so that the channel can validate the invitation using the token. If the allowed node is present and the invitee is not matched against any item, the channel MUST add the invitee to the allowed node as part of the join.

The invitee MAY send an acknowledgement back to the inviter, noting the status of the invitation. This is encoded as an <invitation-ack/> child element of <message/> element. The <invitation-ack/> element is qualified by the 'urn:xmpp:mix:1' namespace. The <invitation-ack/> has an <invitation/> child element that encodes the invitation being acknowledged and a <value/> child element to encode the acknowledgement value. <value/> has the following values:

• 'Joined': The invitee has joined the channel.
• 'Declined': The invitee is not taking up the invitation.
• 'Acknowledged': The invitation is acknowledged, without information on action taken or planned.

Private Messages are used to provide communication with another channel participant through the MIX channel, without the initiating user knowing the real JID of the channel participant. A channel MAY support use of Private Messages. Private messages are standard XMPP messages and MUST NOT be groupchat. A message goes through a number of stages with different addressing. This is set out in the following table.

Private Messages MAY be archived using MAM by the XMPP servers associated with initiator and responder. Private Messages MUST NOT be archived by the MIX channel.

A client MAY request the vCard of a channel participant by sending a request through the channel. The MIX channel MAY pass this request on or MAY block it. vCard requests MAY use &xep0054; (vcard-temp) or &xep0292; (vCard4 over XMPP). The MIX channel does not process the vCard requests, but simply relays them on to real bare JID of the target. A MIX service MAY choose to relay one or both protocols. Where a MIX service relays one or both of these protocols, each protocol relayed MUST be advertised as a feature of the MIX service. In the following example, using vcard-temp, the requesting client sends a message to the bare proxy JID of the channel participant for which the vCard is desired.

The MIX channel MAY pass on the vCard request or MAY reject with an error, dependent on channel policy. The MIX service will then address the vCard request to the user's server (using bare JID) using a full proxy JID to hide the requester.

The user's server, on behalf of the user, MAY send a response or reject with an error. The user's server will send the vCard back to the channel.

The MIX channel will then send the vCard response to the requesting client on behalf of the client sending the response.

Messages sent to participants MUST be archived by both the MIX channel and by the user's server. This MAY include presence messages. Correct MIX operation relies on messages being archived.

Clients will retrieve MIX messages using standard MAM protocol from the user's archive. The MAM query will filter based on the channel JID to enable access to messages from a given channel. This gives the user a simple mechanism to access all messages sent to the channel. MAM can be used to retrieve older messages that have not been cached by the client.

Messages can also be retrieved from the channel by addressing MAM queries to the channel JID. This will behave like a standard MAM archive. This can be useful for administrators to access archived messages. It can also be useful for new channel participants to access the historical archives.

A MIX Channel MAY use MAM to archive nodes other than message nodes. Clients with rights to access these archives MAY use MAM to do this, specifying the PubSub node in the MAM query addressed to the channel.

MIX does not standardize an access control model for creating and deleting MIX channels. The choice is left to the MIX implementer, and could be a very simple or complex approach. A client can determine if it has permission to create a channel on a MIX service, which MAY be used to control options presented to the user. This is achieved by a disco command on the MIX service. If the 'urn:xmpp:mix:1#create-channel' feature is returned, the user is able to create a channel.

A client creates a channel by sending a simple request to the MIX service. A channel MAY be created with default parameters, as shown in the following example. The result MUST include the name of the channel which MUST match the channel name in the request (if present). The create is encoded as a <create/> child element of <iq/> element. The <create/> is qualified by the 'urn:xmpp:mix:1' namespace. The <create/> element MUST have a 'channel' attribute to specify the channel name. This attribute specifies the value that will be used in the LHS of the JID for the MIX channel.

The client MAY also include parameters in &xep0004; format as part of channel creation. If the client wishes to inspect configuration, channel administration procedures SHOULD be used.

Channels MAY be created for ad hoc use between a set of users. Channels of this nature will have channel name created by the server and will not be addressable or discoverable. Here a channel is created without specifying the channel name. Parameters for the channel MAY also be specified.

A common use case for an ad hoc channel is where two users are engaged in a 1:1 chat and wish to broaden the discussion. Prior to bringing more users into a channel, using standard invitation process, there is a need to create and populate a channel. The first step is for one of the two users to create an ad hoc channel, as described in the previous section. The other user will then be invited, and can switch to the new channel.

It can also be useful to share some or all of the messages from the 1:1 discussion into the new channel. The mechanism to do this is to forward messages to be shared in the MUC using &xep0297;. A body SHOULD NOT be used in the outer message. Sharing history is optional. If history is shared, it MUST be done by the user creating the channel before the other user is invited. Any other use of forwarded messages MUST be treated as a member of the MUC forwarding a message to the channel and MUST NOT be treated as history sharing.

There are a number of security considerations with use of MUC history. There may be sensitive information in the 1:1 MUC history, and the user sharing this history should ensure that none of this is sensitive, prior to sharing in this way. The user creating the MUC has potential to inject history messages which were not part of the history. It is recommended that the second user joining the MUC to validate that the messages match the history and to take appropriate action if they do not.

MIX channels are always explicitly destroyed by an owner of the channel or by a server operator. There is no concept of temporary channel, equivalent to &xep0045; temporary room which is automatically destroyed by the server when the users leave. However, channels MAY be configured with an explicit lifetime, after which the channel MUST be removed by the MIX service. Where a channel is created for ad hoc use, it MAY be desirable to keep the channel for history reference or for re-use by the same set of users. Note that the owner of the channel does not need to have presence registered in the channel in order to destroy it.

The destroy operation is encoded as a <destroy/> child element of an <iq/> element. The <destroy/> element is qualified by the 'urn:xmpp:mix:1' namespace. The <destroy/> element MUST have a 'channel' attribute to specify the channel to be destroyed. A client destroys a channel using a simple set operation, as shown in the following example.

A server MUST destroy a channel that has exceeded its specified explicit lifetime. Servers MAY destroy channels which have no participants and/or presence according to local policy. There will often be good reasons to not destroy rooms in these scenarios, in particular to facilitate channel re-use and history access.

Authorized users, typically owners and sometimes administrators, MAY modify the channel information. The client MAY issue a pubsub get command to obtain a form that will facilitate update of the information node. The values in the form show current values, which be defaults or MAY have been explicitly set. In the following example, the channel name was previously set, but other values were not.

Updating the information node is done using a pubsub set command. The MIX channel MUST update the fields with values provided, leaving other fields unchanged. The result returns the id used in the information node item, which is the date/time of the modification.

Channel owners are allowed to modify the channel configuration. The client MAY issue a pubsub get command to obtain a form that will facilitate update of the configuration node. Other clients MAY be authorized to use this command to see the channel configuration, but only owners MAY update the configuration. The values in the form show current values, which MAY be defaults or MAY have been explicitly set. The following example shows a short form returned to illustrate the syntax. A typical configuration form will be much larger with many fields. Modifying channel configuration is done directly by a client.

Updating the information node is done using a pubsub set command. The MIX channel MUST update the fields with values provided, leaving other fields unchanged. The result returns the id used in the configuration node item, which is the date/time of the modification.

Owners and Administrators are allowed to control which users can participate in a channel by use of Allowed and Banned lists using PubSub. These operations follow &xep0060; which sets out detailed protocol use and error handling. Allowed and Banned lists MAY be read by PubSub get of the Banned and Allowed Nodes. This operation MAY be used by users as controlled by 'Allowed Node Subscription' and 'Banned Node Subscription' configuration node options (default Administrators).

JIDs can be added to the Allowed and Banned nodes by a pubsub set command. This is used to add one item to a node.

JIDs can be removed from the Allowed and Banned nodes by pubsub retract command.

When the MIX channel adds a JID to the banned node, other nodes in the MIX channel will be appropriately updated to reflect this change. In particular, the participants nodes and presence nodes will be updated to remove matching JIDs. This will have the effect of immediately removing the user from the channel. For this reason, there is no requirement to have the "kick" functionality of MUC, as this is achieved by banning the user.