The age old challenge with information technology has been machine understanding of information – or rather, the ability to put information about real-world events and content in a model where policies (rules, logic and so on) can act on it.
To boil it down to its simplest level, increase the footprint of information that’s in a model so that it can be processed by an application, and you increase your ability to act on it. Use an information model that’s been proven to work for a given set of processes and assuming you’ve picked the right model you’ve got good processing capability. Enter ERP.
But what about the world we all use to interact with each other – the Web?
In theory, if all the information in the world was in a standard data model, any IT system could act on it. But this is an intractable problem because models are always about predicting something and organisations always evolve and no-one’s figured out how to predict people yet! Hard enough for a medium sized enterprise never mind the world. So enter EAI, MDM and SOA – making us more agile to react to the world.
Now add a world where processes and the information that supports them are extended to customers and citizens through IT – i.e. not under the full ownership of the Enterprise, and it gets even more interesting. Enter Web 1.0, Web 2.0, Software as a Service and Managed Service Provision.
Most people understand the concept of a database – in fact the language has become so pervasive that sometimes the original concept is lost (‘we need a new customer database’ when in fact that’s a solution and the problem might require something very different to solve it).
With a database, we model the data for the purpose of storing it in our database which our application can process. Any data we work with from another party has to be sent or transformed into our data model otherwise our applications can’t work with it. Fair enough.
The semantic Web takes the opposite approach. It starts with the concept of ‘self-describing data’ – i.e. that the data should be understandable by _any_ IT system not just the ‘owning’ application. A way to think about it is to imagine a Web page where the data is marked up in such a way that any system can process it, instead of it taking a person viewing the Web page to be able to understand it.
I think it’s worth reminding ourselves of the problem in its simplest form because I think this year we’re going to see some serious commercial applications of IT to address to it – specifically, the approaches defined by the semantic Web.
The major package and infrastructure vendors already have technology available today which works with some of the semantic Web concepts and standards or have plans to release commercial technology shortly.
But what’s really going to make the shift are people adopting the concepts of information being designed for use by any system and organisations taking market advantage from it. Already some major retail banking, trading and government organisations are starting to deploy the concept.
Patrick Sheehnan, a partner at 3i, says ‘… the semantic Web will be profound. In time, it will be as obvious as the Web seems to us today’.
But will it? I believe so but we have a control hurdle to overcome! With the concept of a database comes the concept of control, corporate governance and industrialisation. Works well for a high-walled and gated Enterprise but the Web has changed business and society and the world isn’t like that today.
Many organisations understandably still hold on to the concept of ‘internal’ information and ‘external’ information, yet the majority of the processes require interaction between both internal and external people and IT.
Can we let go and lose this distinction – i.e. view information as an enabler for any person or application (with appropriate security)?
An organisation that can genuinely make its information resources available to the world – its own enterprise and its web of interactions – perhaps counter-intuitively will exponentially increase its own information processing power.
ERP where E is the world? Given the mass-user adoption of Web 2.0 I think 2007 could see some major organisations seriously investing in their semantic infrastructure – externalising information – to complement the investment in externalising services and processes through SOA.