I shuddered when I read “…machine-understandable way” in Linked Science Core Vocabulary Specification (revision 0.91).
You could say that cash registers “understand” economics because they add and subtract. Machines carry out rote instructions, with no more “understanding” than a bag of hammers.*
As marketing speak the claim of “machine-understandable” is understandable but dangerous. What if a human eader disagrees with the machine’s conclusion/result? “Machine-understandable” imbues the machine with an understanding it lacks. And it has far more access to information than any human.
I’m sorry, if you disagree with the machine you must be wrong.
If the machine gives a result, who will have the courage to contradict it? Consider something as simple as moronic search results. Airline security personnel who detain four and five year olds with names similar to known terrorists won’t contradict obvious bad search results. How much more will they credit results bases on a “machine-understandable” interchange of data?
Let’s be clear about who has understanding, that would be human users (I leave the potential for sentient aliens to one side), and the responsibility to evaluate results. Whether from a vending machine, cash register or OWL reasoner.
* Yes, I am aware of adaptive programs that reach results not foreseeable by their programmers. But, the adaptive process wasn’t designed by the machine. Rather a human author, someone with “understanding” wrote the relevant instructions.