From Semantic Portal Wiki

General Description

There are various mediums of publishing and getting "information" on the web - blogs (text, images, audio and video), wikipedias, news websites, corporate websites etc. One publication usually refers to information at other spaces, in the interlinked information space of the web. Then there are aggregators (in form of web pages, desktop client and browser extensions) that provide collage of information collected from the above sources of information - web search, news.google.com, planetrdf.com, headsup etc. Mechanisms for associating trust with information, depending on various factors including its source, and utilizing them while presenting it, are being investigated extensively both in academia (Jennifer Golbeck) and in the industry (IBM's web services trust language).

<What constitutes information flow? examples>.

Many factors affect the flow of information on the Web and have various impacts. As our first use case we will consider the blogosphere - as considerable amount of research in the past 5 years has been centered around it. Various models have been proposed to capture the information flow. The common denominator in all those models, it seems, is the factors - number of in-links from other blogs (posts), number of posts, number of out-links from the blog (post) <TODO - add more, refine>- that are considered fundamental to differentiate flows in different scenarios. These factors affect the topology of blog space. In our first scenario we capture these changes in topology and hope to explain and predict information disseminations of the past and in the future, respectively. Addition of a new post to the blog space and addition of links to the old posts from this blog - are some of the temporal events that would be captured using the event Ontology.

All the events in relation to a blog space under evaluation, that change the topological space constitute a process. Rather than collecting events directly under a process, they can be considered part of different states, and change in states could constitute a process. The processes can be characterized by the type of change (from star to chain for example), the rate of change, size of change etc. This can help answering questions such as "characterize the flow of information X published at time t1, in space A, that reached space B at time t2, where A and B are (dis-)connected spaces", or "explain why a particular blog became smash hit in community C but not in community D". None of these questions can be answered completely or correctly, but relevant process and constituting events can be returned as answers to such queries.