Integration - a way to provide better navigation

Integration is one of the key aspects which makes things work on the web. In the field of navigation we consider following levels of content integration: 

• Statically linked content - static links
• Dynamically linked content - dynamic links or page chunks like weather forecast widgets
• Semantically linked content with regard to its' semantics - like Wikify!, which links text to Wikipedia articles

During the evolution of web, the static linking was the first level of content linking, which is simple, widely used and sufficient for many purposes. It provides almost absolute control over the integration - unless the target content becomes irrelevant or unreachable. The expansion of databases usage enabled dynamic content linking. This is a big improvement of integration, but not as big as it would seem. Relevant content is still integrated by humans, but the results of integration are presented to the user dynamically. Imagine news portals - regular news portal provides the "Similar articles" section, which enables a shortcut to older or more recent related articles. Who created relations between these articles? There may be some recommendation systems helping out, but in the end, it is journalist or editor who sets them. So the relations have moved from pages to databases, what makes their use more flexible and manageable.

Now what the semantic content linking adds to this stack? It lets page linking and content linking do what they are good at - creating relations that can be determined at the time when content is created, and it introduces a new concepts - context dependance and knowledge linking. Knowledge linking enables access to the content with regard to what it "means". It is not linking of pages based on human made relations, but linking of content using human made knowledge (it is likely that computers will be able to gather knowledge too, for now they can make inference from formally represented facts, but are not that good in identifying meaningful new facts from unstructured text). In the field of web navigation, the context dependance is important. The context is not the page read by the user - it is the entities, facts and finally knowledge contained in the text of the page. The key is to make computer "understand" the context and identify it correctly. When this is accomplished, semantic content linking can open door to more simple orientation on the web. In fact, that is what we long for. We want simple access to the content that would otherwise be hardly discovered, and we want it as fast as possible.

There are formal issues that enable knowledge integration (ontologies, linked data, ...) which will be discussed in later articles. In the next one, we will talk about integration of different content resources available on the web. When mixed together in the right proportion with some extra processing involved, they open unexplored paths right from the page that you read.

See you soon ;)