A lot of activities around Linking Open Data (“LOD”) and the associated data sets which are nicely visualised as a “cloud” are going on for quite a while now. It is exciting to see how the rather academic “Semantic Web” and all the work which is associated with this disruptive technology can be transformed now into real business use cases.
What I have obeyed in the last few months, especially in business communities, is the following:
- “Linked Data” sounds interesting for the business people because the phrase creates a lot of associations in a second or two; also the database crowd seems to be attracted by this web-based approach of data integration
- “Web of Data” is somehow misleading because many people think that this will be a new web which replaces something else. Same story with the “Semantic Web”
- “Linking Open Data” sounds dangerous to many companies
For insiders it is clear, that the “openness” of data, especially in commercial settings, can be controlled and has to be controlled in many cases. Which means, it can be one-way or mutual. In some use cases data from companies will be put into the cloud, and can be opened up for many purposes, in other use cases it will stay inside the boundaries. In other scenarios only (open) data from the web will be consumed and linked with corporate data, but no data will be exposed to the world (except the fact, that data was consumed by an entity).
And of course: In many other occasions datasets and repositories will be opened up partly depending on the CCs and the underlying privacy regulations one wants to use.
This makes clear that LOD / Linking Open Data is just one detail of a bigger picture. Since companies (and governments) play a crucial role to develop the whole infrastructure, we need to draw a new picture:
I´ll be happy to have a lively discussion about this topic also at the first Linked Data Camp in Vienna.
Pingback: The Semantic Puzzle | Linked Data Flows: A new picture to illustrate the “openness” we mean
You might complete this picture by adding the view that also opening data sources inside an organization (w/ usage of linked data priciples) provides value…even if this is not directly usable in the internet, but ‘only’ in an intranet.
So semantic web standards (combined with applied linked data principles) can be an integration layer.
Today there is not enough tool support for that: most the middleware products imply that the system integrator “knows” the connections between the data, but it is very seldom described in standardized, machine-readable documents.
Key for integration usage of linked data principles are from my point of view:
– data synchronization tools (e.g. to create batch jobs to update RDF-stores – because one-time imports are not always the solution) and
– more sophisticated search technologies (which answer questions like “give me all contacts in customer organizations in healthcare sector in Spain, with which we did business in the last three years” . This is easy if you are a small company in healthcare sector in Spain, but gets complicated if there are different CRM systems in place)
Thanks Daniel, yes indeed: RDF as a technology for data integration only inside the firewall is also a possible application scenario (which is indicated in the picture with the arrow to the upper right side).
Some tool vendors like OpenLink Software (Virtuoso) are able to provide stable middleware-systems for this kind of application.
Pingback: Will the Semantic Web have a gender? « Fredzimny's Blog