Yeah I reckon so! I think the big issue I have with them though is scaling them to full language specs - the current inference rule notation works for small stuff, but can get unmanageable quickly.
I was thinking it would be neat to have some text notation to define the rules, and then a way of presenting them that implementors can use - perhaps with a way of switching between a diagram version, generated 'prose', and more traditional maths notation.
With clickable links to jump to definitions and usages, ways of grouping based on judgement vs syntactic construct, and an index of rules, judgements and syntax.