What I've later come to realize is that it was quite a hacky demonstation project with a non-scalable RDB model that proves that:
As engineers, we love to create our own fancy systems that solve one particular task. But how will it scale? How does it fit with sorrounding data processing tasks in the project?
In my research I realised that while creating cool tools is awesome,
the important part is to establish a common language
and shared understanding
...but as our industry involves so many different stakeholders, a global common understanding is unattainable!
The very intention of BOT is to provide a minimal, extendable common core.
How can boundary conditions for walls be described in a semantic data model so:
A set of rules were defined
A demo model was created
Rather simple: "What's the wall's max height?"
For the remaining rules, instead of describing properties directly on the walls, they were inherited to wall surfaces based on space demands
Something heavy is mounted on the wall -> it needs a reinforced surface!
A quiet room and a loud room are separated by this wall -> it needs acoustic insulation!
A heavy door is carried by the wall -> it needs structural reinforcement!
Intelligent systems can infer the implicit knowledge from the bits and pieces that we feed it
Incremential reasoning
The technology is maturing - all the bits and pieces are already there
Go and create something meaningful with it! 🚀🚀🚀
Questions? Thoughts? Ideas? Let's discuss them over a ☕