Web 3.0 vs the Semantic Web

2009-09-21 Desk

What's in a name? The term Web 2.0 was never really intended to be a formal definition. In relation to the first generation of static HTML websites, one could see rather intuitively the direction that web 2.0 was taking. Tim O'Reilly widely cited mememap and definition helped a lot.

Since then, we have come to know Web 2.0 applications as online architectures supporting to different degrees open, interconnecting functionalities, more popularly known as ‘social networks’. The transition to Web 3.0 is even less clear cut, and using the term Web 3.0 interchangeably with the term 'semantic web' makes things even more confusing. Let's take a look at what these expressions really mean, and let's see if we an learn how to use them appropriately.

It's all about relations

The capabilities of what has become known as Web 2.0 are provided by 'relational' technologies, coupled with synchronous JavaScript and XML (aka AJAX) arranged to provide 'interconnectivity' of data and information that they contain, made appealing by intuitive user interfaces. In the relational model, data is arranged logically and represented as unique relations. The sorting and relating of cata in tables makes it possible to leverage (at least some of) the inherent semantic associations that can exist in cata expressed using natural language. Simple queries can be performed to retrieve attribute that pertain to an object or entity has, or what relation there is between objects and attributes. Typical queries are based on operations that correspond to union, intersection, difference, product from mathematical set theory, and selection, projection, natural join, relational division, which are proper to relational algebra as devised by Codd in the seventies. These operations can be performed provided each value in the database is unique. Exact advanced algorithms capable of more precise or faster queries tend to be 'proprietary' and constitute the unique selling point for many software applications. When creating an account and login on web 2.0 applications users create a table (user profile) then the application ill typically store every bit of information entered, and it will even store dynamically as a relational table certain 'behaviours'. Querying and manipulating such information is a trivial task for relational operators. That's why it's possible to see what our friends are doing, what applications and topic may be relevant to us, etc: because all the information is in the same database (the application database). When it comes to the open web, however, querying and manipulating data which is stored in different, closed databases, cannot be done without merging the respective databases. Therefore RDF and OWL (the current semantic web standards) were devised to represent the data outside the database presumably on the web, but on intranet works), the preferred terminology in use is to 'expose' and to 'link'. However semantic queries over 'exposed' data sets still rely on the relational principles, therefore the importance of unique values, such as 'URIs' and 'URLs' and 'URNs' etc. Useful applications can also leverage the semantic capabilities as explained above to integrate heterogeneous data repositories within the same organisation. Web 2.0 and semantic web discussed without understanding the role of relational principles can result in terrible confusion. A good example of limited understanding of semantic web and web 2.0 altogether is provided here . Time to start throwing some old furniture (especially Chairs) out of the window perhaps.

Beware of the Happy Semantic Bandwagon

Confusion has its costs, and it is likely to generate scores of profits for those who are in the business of Selling 'solutions'. They have no intention of shedding clarity, dispel myths about the semantic web, otherwise they would make no money. When considering migrating to semantic architectures, make sure you do not necessarily listen to every researcher, faculty, consortia and their industry partners who are all trying to capitalize on the general lack of understanding of semantics. After over a decade of ignoring or even criticising and attacking the semantic web, many are trying to capitalize on its renewed charm. Semantic capabilities, with a few exceptions such as skills acquisition and high end software packages which are worth their salt, are intrinsically about exposing and linking data. That is likely to require a lot of rethinking, of the kind that money cannot buy. Beware of semantic vapourware currently being aggressively pushed to markets worldwide as semantic technology.

Bookmark/Search this post with:

Comments

Post new comment