Talk:YAMLish formal syntax

From Test Anything Protocol

Redefining the whole syntax worries me greatly. Now we can't be sure if we support a true subset of YAML or if we've made modifications.

I would rather see YAMLish defined as "the YAML spec minus these parts". Subtractive. This guarantees that we support a subset and also guarantees that existing YAML parsers will handle YAML. It also means we won't forget anything really important.

--Schwern 00:14, 20 March 2007 (GMT)

Yeah, you're probably right. It's not exactly the most rigarous specification anyway. I was concerned about our perlcentricity but your splendid idea of distributing acceptance tests as YAML documents addresses that far bettter. I'll orphan the formal syntax page until I decide whether anything useful can be salvaged from it.

--AndyArmstrong 01:05, 20 March 2007 (GMT)

I think supporting the inline styles, { foo: 42, bar: 23 } and [1,2,3] will be useful for brevity of output. Especially for lists.

--Schwern 00:15, 20 March 2007 (GMT)

That's more or less JSON then :)

I'm cool with that. Of course implementing it moves YAMLish inexorably in the direction of YAML...

--AndyArmstrong 01:07, 20 March 2007 (GMT)