Create rubrics for testing and metrics
kerrizor opened this issue · comments
We could use a number sample classes that contain latent objects in order to gauge how effective automated sniffing is. Ideally these are both obvious and not-so-obvious examples of latent objects. I'm not entirely sure we could objectively quantify the objectness of a recurring concept, it'd at least serve as a starting place for discussion.
Would you want those as a PR, discussion here, or a wiki page?
You can also make your README executable so that your docs demonstrate usage...
I'd love them in any of those formats, to be honest. :)