Problem

From the consumer’s point of view, with everyone having its own channel, how are people going to find what they want amidst all this avalanche of information? The Web as it is today, as complicated as it might be to publish something, seems already overcrowded with its billions of interconnected documents. How then are people ever going to find what they want when technology easies the publishing process even further and opens hundreds of millions of hidden worlds all filled with rich personal media?

From the producer’s point of view, how will he or she broadcast his or her channel so that it reaches those to whom it is relevant without interfering with those to whom it is not? How will his or her channel stand apart from the so many others that exist? How will he or she get his or her idea across and be heard in this open and democratic medium, where everyone has gained the right and opportunity to have a voice?

Proposal

ISS (Instant Syndicating Standards) addresses these challenges by unleashing the great deal of untapped potential of the collective intelligence to organize information. It exploits the bottom-up characteristics of successful technologies and finds the perfect balance.

The idea is to connect people with news and information that matters to them the most by letting people themselves express what matters to them at an individual level. Each individual connects with their own personal social network to receive and disseminate information. This challenges the top-down model of information-sharing and gives place to a bottom-up model, where each person has a unique voice and equal opportunity to contribute and benefit. In this way, ISS hopes to bring people closer together to discuss common interests and share information in a more open and democratic manner.

Peer-to-Peer Networks

Peer-to-peer technology refers to a class of applications that take advantage of the resources available at the edges of the Internet. These resources can either be computing power, storage space, content, bandwidth or human presence.

Peer-to-peer applications have proven to be important for many tasks such as distributed computation (e.g. Seti@Home, Genome@Home), content distribution (e.g. e-Mule, BitTorrent) and communication and collaboration (e.g. Jabber/XMPP, Lotus Sametime).

One of the biggest challenges of peer-to-peer applications is the lookup problem, i.e. how to address the resources available in the network. Some lookup techniques are: centralized search (e.g. Napster), flooding (e.g. Gnutella), Distributed Hash Table (e.g. CAN , CHORD), semantic routing (e.g. Edutella, GridVine), social routing (e.g. Bouillon, WiredReach) and publish/subscribe (e.g. Siena, Scribe).

Among the main advantages of peer-to-peer technology are robustness, scalability and easiness of use. Bringing the user's personal computer closer to the heart of the Internet makes the whole sharing process much easier.

References

Social Networks

A social network represents the relationships people and organizations have. Social network analysis studies these relationships and their applications.

Among the classic social network analysis are the small world problem, from Stanley Milgram, the small world phenomenon, from Duncan Watts and Steven Strogatz, and the strength of weak ties, from Granovetter. These studies demonstrate how connected we all are and how important social networks are for the dissemination of information.

Many researchers have studied the impact and the potential social networks have on applications such as e-mail (e.g. Malone et al. with the Information Lens System), mailing lists (e.g. Goldberg et al. with Collaborative Filtering), newsgroups (e.g. Riedl et al. with Grouplens), personal sites (e.g. Kautz et al. with the Hidden Web), search engines (e.g. Brin and Page with PageRank) recommender systems (e.g. Maes et al. with Social Information Filtering), reputation systems (e.g. Resnick et al. with Reputation Systems), peer-to-peer networks (e.g. Grassman et al. with Social Networks in Peer-to-Peer), instant messaging (e.g. Smith with I.M. as a Scale-Free Network) and blogs (e.g. Kumar et al. with Structure and Evolution of Blogspace).

Web applications that help people organize their social network online are very popular. Some examples are MySpace, Facebook, and Orkut. These applications are walled gardens, but open standards such as FOAF (Friend of a Friend) and XFN (XHTML Friends Network) also exist.

References

Instant Messaging

Instant messaging provides real-time communication between two or more people and often provides presence information as well. Some examples of instant messaging applications are ICQ, AIM, Yahoo Messenger, Microsoft Messenger and Jabber/XMPP.

Recently there has been a great effort towards integrating multimedia such as audio and video into instant messaging. Applications that offer audio are known as VoIP (Voice over IP). Examples include Gizmo, Asterisk, WengoPhone, Skype and Google Talk.

Real-time communication and rich awareness are propitious for informal collaboration and often are part of groupware applications. Some examples of groupwares are Lotus Sametime, Adobe Acrobat Connect and Groove.

There have been several attempts to create a unified standard for instant messaging. The most popular standards are SIP (Session Initiation Protocol) and Jabber/XMPP (eXtensible Messaging and Presence Protocol).

References

Blogging

A blog is a user-generated website that resembles a journal or diary. Some common examples of sites that let users build ther own blog are LiveJournal, Blogger, and WordPress.

A blog is made up of entries that are displayed in a reverse chronological order and that typically consist of: a title, a date, a body, a permalink and tags. Blogs often provide means of interaction and socialization such as commentaries, trackbacks and blogrolls.

There are two main blogging format standards: RSS (Really Simple Syndication) and Atom. Aggregators that support these formats help people organize multiple subscriptions.

Blogs have become the de facto personal channels of the Web. They are very easy to setup and they help people share ideas, photos (flogs), audio (podcasts), video (vlogs), and presence (microblogging).

References

Social Tagging

Social tagging is a bottom-up approach to cataloging information. Traditional cataloging of information is done by experts using taxonomies or ontologies in a top-down fashion. Social tagging, on the other hand, is done by common people. People are free to choose any set of keywords that they see fit to describe a specific information.

Social tagging is a quite recent phenomenon and has being very successful for organizing information. They are used for organizing e-mail (e.g. Gmail), bookmarks (e.g. Del.icio.us), photos (e.g. Flickr), music (e.g. LastFM), video (e.g. YouTube), blog entries (e.g. WordPress), and reference entries (e.g. Wikipedia). Social tagging main advantages are a) the easiness of use, since tags correspond to a flat namespace; and b) the emergent collective intelligence, since the "selfish" effort of individuals generally helps the community as a whole.

References

Content Syndication

Content syndication helps users to publish and be kept aware of any changes to websites, human availability, or anything of interest that changes with time.

Applications today usually follow the pull model, i.e. they apply polling techniques to check for updates. Some issues regarding the pull model are limited scalability and bandwidth usage. To address these issues, some applications follow the push model, i.e. they receive updates automatically. Some examples of push technologies are XMPP, FeedTree, Cometd and SamePlace.

There are two content syndication standards: RSS (Really Simple Syndication) and Atom Syndication Format. The Atom Publishing Protocol also describes how applications may publish updates.

References

ISS Identity

An individual may be identified using a(n):

• URL: for the Web;
• JID: for XMPP.

<div class="vcard">
  <span class="fn n">
     <a class="url" href="http://nick.iss.im">
       <span class="given-name">Nick</span> 
       <span class="family-name">Vidal</span> 
     </a>
  </span>
  <span class="nickname">nickvidal</span>
  <img class="photo"
       src="http://nick.iss.im/wp-content/nick.jpg" />
</div>

ISS Core Formats

ISS Core Formats are:

• TagCloud: describes the tags from an individual;
• TagLink: connects these tags through a trusted network of people.

ISS Core Services

ISS Core Services are:

• Discovering: display tagclouds for each contact;
• Subscribing: subscribe to friend's tags;
• Viewing: view entries from subscriptions;
• Syndicating: syndicate entries to subscribers.

They can be implemented using:

• REST/AtomPub: for web-based applications, and/or
• PubSub/PEP: for IM applications.

Discovering

One of the main features of ISS is the ability to discover friends’ current interests by viewing their tag cloud. A tag cloud is a visual representation of the most popular tags in a given time frame. The Figure below shows a tag cloud for Debbie for September 2006. The most popular tags are Health, Spirituality and Yoga; followed by Dance, Education, Music, Romance and Shopping. To create a tag cloud, it's necessary to compare how many entries there are for each tag in a given time frame. The more entries it has compared to other tags, the bigger it appears.

Chat Window

Besides size, colors also play an important role in tag clouds. The colors of the tags represent the incoming and outgoing flow of information. The blue color is for tags that have an incoming flow (e.g. Nick receives information from Debbie's Friends tag). The red color is for tags that have an outgoing flow (e.g. Nick sends information to Debbie's Spirituality tag). The purple color is for tags that have a mutual flow (e.g. Nick receives and sends information about Art from and to Debbie). Finally, the black color is for tags that have no flow of information (e.g. Nick and Debbie don’t exchange information about Environment – yet).

Let us suppose Nick is interested in receiving entries from Debbie that has been tagged with Environment. All he has to do is to select Debbie’s Environment tag to see the latest entries she has written about the Environment. If he is truly interested in subscribing to Debbie’s Environment tag, then all he has to do is associate Debbie’s Environment tag with one of his tags.

Retrieving Debbie's TagCloud in XMPP:

<message from='debbie@iss.im' to='nick@iss.im'
type='headline' id='discovering'>
<event xmlns='http://jabber.org/protocol/pubsub#event'>
<items node='iss'>
<item>
<iss>
…
<tag name='environment'>
<tagcloud type='syndicated' year='2006'
data='0,0,0,0,0,0,0,0,3'/>
<taglink type='base'
push='xmpp:debbie@iss.im?;node=environment'
pull='http://debbie.iss.im/category/environment'/>
…
</tag>
…
</iss>
</item>
</items>
</event>
<x xmlns='jabber:x:delay' stamp='20060921T21:27:44'/>
</message>

Subscribing

When subscribing to a friend’s tag, the users’ own tag cloud is shown. This tag cloud is shown in the Figure below. The users have to select the tags to which their friend’s tag will be associated. A tag may be associated with one or more tags.

View Window - Subscribing

Let us suppose Nick associates Debbie’s Environment tag with his Science tag. Now every time Debbie publishes something tagged with Environment, these entries are automatically tagged with Science.

Thus, for users to share information, they don’t have necessarily to share the same concepts. All they have to do is to show how their friends’ concepts are associated with their own concepts. This is what tag links do. The Figure below shows Nick’s tag links. Debbie’s Environment tag is now associated with Nick’s Science tag. Since Nick is subscribed to Debbie’s tag, the color of the tag link is blue (incoming link). Outgoing tag links are in red.

Subscribe Window

Subscribing to Debbie’s Environment tag in XMPP:

<iq type='set' from='nick@iss.im'
to='debbie@iss.im' id='subscribing'>
<pubsub xmlns='http://jabber.org/protocol/pubsub'>
<subscribe node='environment' jid='nick@iss.im'/>
</pubsub>
</iq>

Viewing

The figure below shows the View Window. It resembles an aggregator with lots of information shown. Again, colors play an important role. Entries in blue are incoming entries, i.e. entries that have been received. Entries in red are outgoing entries, i.e. entries that have been sent. And entries in purple are entries that have been shared. The star with a smiley icon shows whether the user has liked an entry and decided to share it. If, within a group of subscribers, many people have shared a specific entry, then the three stars to the right will be shining. And if someone has added a comment to the entry, the far left star will be shining. Thus, if all five stars are shining, then this is a very relevant entry within this social network.

View Window

The icons in the top bar are guides to help the user filter entries. If the user wants to filter Science and Technology entries, for example, he or she can select the tag icon, which displays the user’s own tag cloud, from which he or she can select the Science and Technology tags. In the search box it will appear the following string: "tags(Science and Technology)". Other combinations can be done as well.

View Window - Filtering

It is important to remember that even though the entries have been tagged as Environment by Debbie, they are tagged as Science for Nick. But these entries are not published in Nick’s Science node unless they have been shared by Nick himself.

Retrieving entries from Debbie’s Environment tag in XMPP:

<iq type='get' from='nick@iss.im'
to='debbie@iss.im' id='viewing'>
<pubsub xmlns='http://jabber.org/protocol/pubsub'>
<items node='environment'>
</pubsub>
</iq>

Retrieving a specific entry in XMPP:

<iq type='get' from='nick@iss.im'
to='debbie@iss.im' id='viewing2'>
<pubsub xmlns='http://jabber.org/protocol/pubsub'>
<items node='environment'>
<item id='368866411b877c30064a5f62b917cffe'/>
</items>
</pubsub>
</iq>

Syndicating

If Nick wants to share a specific entry, he may do so by tagging the entry appropriately. The Figure below shows that this particular entry is already tagged with Science. He may add new tags by selecting them from his own tag cloud, or he may create new ones by typing them. When he syndicates this entry, he publishes this entry to the nodes of these tags. Everyone subscribed to these nodes will receive a notification.

Syndicate Window - Tagging

Many metadata may be associated with an entry. In the Figure above we can see five metadata associate with this entry. The first metadata indicates that this entry is tagged with Science. The second metadata indicates the original author of the entry. An entry may travel far within the social network, but the original source should always stay intact. The third metadata is the file attached to the entry. This file travels from person to person, computer to computer, peer-to-peer style. In parentheses it’s shown who in Nick’s personal social network has already downloaded the file. Nick can download the file directly from them instead of downloading from the source. Information about the file’s location is published to the Downloaded node. The fourth metadata is comments. Nick can read what comments his friends have about this particular entry. Comments are published to the Commented node. And finally, the fifth metadata is the rating of this entry. Three of Nick’s friends that are in his incoming tag link for this tag have shared this entry. If Nick decides to share this entry as well, the three stars will be shining. Shared notifications are published to the Shared node.

Syndicating an entry in XMPP:

<iq type='set' from='nick@iss.im'
to='pubsub.iss.im' id='syndicating'>
<pubsub xmlns='http://jabber.org/protocol/pubsub'>
<publish node='science'>
<item id='ae890ac52d0df67ed7cfdf51b644e901'>
<entry xmlns='http://www.w3.org/2005/Atom'>
<title>Tips for a Better World</title>
<author>
<name>Debbie Soares</name>
<link jid='debbie@iss.im'
id='368866411b877c30064a5f62b917cffe'/>
</author>
...
</entry>
</item>
</publish>
</pubsub>
</iq>

ISS Extended Formats

ISS Extended Formats are:

• Read
• Shared
• Downloaded
• Commented
• Searched
• Matched

ISS Extended Services

ISS Extended Services are:

• Downloading
• Commenting
• Rating
• New Tag Notifying
• New Subscription Notifying
• New Friend Notifying
• New Identity Notifying
• Auto Syndicating
• Searching

ISS Policies

Some policies to control the quality of information are:

• Incoming information only comes from one-degree apart friends through subscriptions;
• Friendship requests only come from people x-degrees apart, where x is set by the user (a graph should be created beforehand);
• Friendship requests are viewed as a special channel, where the user can easily see the person's tagcloud, how many degrees separate them and how much they are connected;
• The user may decide to provide services only to people x-degrees apart. This is particularly interesting for resource-demanding services such as Searching, for example.

Centralized and Decentralized Networks

ISS uses peer-to-peer technology and finds the perfect balance between centralized and decentralized networks. The robustness, scalability and easiness of use of the decentralized model are its greatest advantages. Yet, there is also the simplicity of the centralized model. The flexibility of Jabber/XMPP technology, on which ISS is built upon, let’s users enjoy the best of both models.

Anonymous and Trusted Relationships

ISS uses social networks technology and finds the perfect balance between anonymous and trusted relationships. The cascading of trustful social networks works as a world wide distributed recommender system perfectly tuned to output a very personalized journal. This trustful network filters out irrelevant information, while still letting good information pass through. Users receive not only information from their personal social network, but from people they don’t know as well (as long as it is recommended by their friends).

Asynchronous and Synchronous Communication

ISS uses instant messaging technology and finds the perfect balance between asynchronous and synchronous communication. Instant Syndicating let’s users share and discuss information in real-time. Yet, users can catch-up with updates in their own time.

Open and Structured Knowledge Management

ISS uses blogging technology and finds the perfect balance between open and structured knowledge management. The well structured chronologically organized entries are ideal to be passed on from person to person. Yet, the entries are still flexible enough to be used in many creative ways. For example, a channel could be created by a user to help his/her friends to be kept aware of new contacts he/she adds to his/her contact list. The system can publish information about the new contacts automatically and broadcast it to the subscribers.

Expert and User Metadata Creation

ISS uses social tagging technology and finds the perfect balance between expert and user metadata creation. Every user creates their own tags. To share information, users don’t have to share the same tags. All they have to do is to show how tags are associated.

Pull and Push Technology

ISS uses content syndication technology and finds the perfect balance between pull and push technology. Users won’t get information they don’t want. But they will still be able to discover new things. With ISS, users discover by seeing tag clouds first; and if they do find something interesting, then they can subscribe to it.

Contributions

• Integrating different technologies in a seamless framework;
• Creating a simple yet elegant workflow for information-sharing;
• Creating the concept of taglinks;
• Describing tagclouds and taglinks in XML;
• Creating a protocol for syndicating tagclouds and taglinks over HTTP;
• Creating a protocol for syndicating tagclouds and taglinks over XMPP;
• Exploring tagclouds to provide a better user interface for self-awareness, peer-awareness, information subscribing, and information filtering.

Future Work

• Promote ISS;
• Analyse the impact ISS has on information-sharing;
• Draw some nice graphs demonstrating this impact.