Should the implementations be scored?

[ohler55]

Should the implementations be scored for a quick glance summary?

It could be as simple as making the score the sum of the results of each test. Something like:

✓ The result of this implementation matches the consensus of results. 1.0 points ✗ The result does not match the consensus. 0.0 points ➚ and ➘: no clear consensus amongst the implementations, but ➚ indicates a majority (and possible future consensus). majority gets 0.7 points, minority get 0.3 points e The implementation failed executing the query. 0.0 points

Comments?

[ohler55]

I guess there is no interest in this. You are welcome to close.

[cburgmer]

Hey, thanks for checking back. Happy to keep this for a while, as I remember this question coming up before!

[bwl21]

I would love to see this happen.

[ohler55]

@cburgmer if you think it is worth doing I can put together a PR for the scoring. What do you think?

[glyn]

I like the idea of scoring against the consensus. In the fulness of time, supposing a standard is defined, perhaps we could also score against the reference implementation.

[danielaparker]

If I could add one further suggestion, of the 41 implementations, two early ones have had disproportionate impact on the JSONPath ecosystem, and to some extent account for the divergent evolution of later ones - Jayway Java and Goessner Javascript. It would be interesting to see scoring against these two important ones as well.