[Graham Meaden]

On your third key difference (P.5), you are absolutely right, the connections become so important.  Also, it’s not just the built asset that needs to be represented: it must include the taxonomical representation of safety, operational and environmental risks and how they interact with the built environment.  In the Common Information Model we have been building for Intelligent Infrastructure programme at Network Rail we have included provision for modelling how these types of influence generically apply to classes of asset and how they may influence individual assets over time to alter asset performance and capability.

[Graham Meaden]

SOmething that jumps out for me on the first page…

“A Digital Twin is a digital representation of a physical thing (and the logic of its operation) that one can query”

You could extend with “…and exert control” as I assume remote control would be a longer term objective?  If nothing else software-based devices and equipment will require periodic software updates.

[Graham Meaden]

Ian, we have a working MVP called Graphshare trying to address this. Let’s catch-up.

[Graham Meaden]

Ian, the problem with so many tools is that they forget that it is the audience needs that should drive the content and what and how it is presented. I believe we need to separate the building of models from the consumption of models.

I would observe that in too many cases software products present data in graphs “because they can” rather than because it helps the consumer make a decision or gain understanding.  Many software products presenting graphs fail to deliver value because the data presented is unworkable, unusable, overwhelming or all of the above.  In varying degrees tools can present:

• too many nodes

• too many links

• too unpredictable

• too general a set of nodes

• too general a set of links

• not enough information about nodes or links when inspected

• too much information about nodes or links when inspected

• nodes that cannot be navigated from

This problem is summed up in the phrase “you can’t see the wood for the trees”.  To take the example above and in the Ontopop video, what is the tool for: building model, exploring a model or supporting a repeatable process? The challenge in my mind is how do we practically bound a graph query and deliver the right information to the user first time on a repeatable basis?

We should leverage the ISO42010 approach and ensure we focus on Stakeholders, Concerns and separate Viewpoints and Views from the underlying Model.