WikiSim is an Executable Knowledge Graph as a Public Wiki

With WikiSim I am building a public computable knowledge graph to support understanding which actions and policies (government or otherwise) are possible from those which are just wishful thinking.

This will mean that when someone makes a claim such as

"To achieve X it will cost £Y"

or

"We can only do X, we can't do Z to tackle problem Y."

then those values of XYZ can be - should be - derived from a whole bunch of other values and that's a knowledge graph. e.g. it might look something like:

Disagreement

That means that if someone disagrees with that claim, then instead of being vague, they can now be specific about what they disagree with. For example if there was a knowledge graph like this:

... it is clearly wrong in a couple of places:

1. Y should be calculated by multiplying A and B, not adding them
2. the "Cost of B in €" node is not wrong but it's being used in the wrong place and instead the node - or data component - that should be referenced is "Cost of B in £".

Disagreement Resolution

With this more detailed view of the knowledge and disagreements it would then be trivial to either

• fix the problems either directly in the public wiki or
• by proposing an alternative graph

WikiSim should make it easy to compare the two graphs to understand where exactly the disagreement is.

Validating from Source

When these knowledge graphs are used to power a visualisaton, or simulation / serious game, then individual numbers and properties that emerge can be directly traced back to the individual components of the knowledge graph.

This means it is much easier to show how the claim is derived from data and assumptions and this should make it easier for everyone to understand and verify the claim, i.e. to easily see the data it is derived from, its context, the calculations performed on that data, and understand the different assumptions behind those data and calculations.

A Real World Example

In 2009 David MacKay published his book "Sustainable Energy - without the hot air" and whilst loading it into WikiSim to power a new national energy game I realised it had a minor mistake in it. This shows up as a different between the numbers in the energy budget "stacks" from the book and those that are now in the Wiki:¹

You'll see the difference also shows up in the totals presented at the bottom of each stack.

This is already a great improvement over existing visualisations which would not easily show the difference between two different perspectives on the world. But we can do better than this.

If we want to know where exactly the difference is we can view the underlying knowledge graph and drill down to the exact data component / nodes which are different:²

I'm hoping that this view might one day be available for many more discussions to help people more quickly pin point where exactly disagreements lie and to resolve them efficiently either through:

1. agreement i.e. direct convergence
2. reframing disagreements as a question / experiment to run to find out the answer.
3. agreeing to disagree i.e. accepting that there are multiple options and that any plan needs to be flexible enough to accommodate them all.