Glenn Jackman

The new tests take different inputs, so it seems like they should have different property names.

Nobody's commented on this so far: my take: The README describes it like > " The 'property' is a string in lowerCamelCase identifying ", > " the type of test,...

Seems to be a clear consensus. Thanks for the opinions folks. I don't currently have the gumption to carry this forward.

I think we also need a test where the validations succeed initially, but fail further down the tree: This input is the existing "Tree with many items" data with the...

This is probably more apropos for the configlet repository. Perhaps place a `.keep.docs` file in the exercise's `.docs` directory. That can inform configlet to not sync the exercise's docs.

The [micro-blog](https://github.com/exercism/problem-specifications/blob/master/exercises/micro-blog/description.md) exercise deals with handling unicode characters, so there's a bit of overlap with this suggestion.

I don't understand why you're parsing html (with regex!) at all. Aren't you inserting the opening `ul` tag when you start a list?

As you've discovered, that's not a reliable way to process the file. I'd recommend adding the start/end ul tags when needed while you're detecting list items.

In the Tcl implementation, I did ignore the string marshalling aspect, and used native dictionaries for the input and output data type.

Please see http://forum.exercism.org/t/maintenance-help-needed-on-groovy-track/12434 before merging this PR.