Our protocol theory applied to job hunting

Take the basic idea, every employer has decision trees, and the job seeker has one.   We can run the convolution, but all the job descriptions are an unordered set. But we have agreed sector semantics,  both sides understand sectors. And wage ranges, even as conditionals, have some segmentation. 

So, with a little work, we call the job interview a protocol, if the protocol runs to completion, all components, each a decision tree, reach the completion state; then the parties meet for an interview.   Convert Watson into a protocol machine,it can decompose and index on the protocol graph itself.. Then perform distance measures  between protocol graphs. Stepping toward singularity 2.0.