Chris Llanwarne

@patmagee if we do this, a more intuitive json input might be a good idea too- do you have any thoughts on what that might look like? Right now it's:...

@patmagee this was an oversight in Cromwell's implementation. I guess the fact that it lasted this long means it's a bit of an edge case but it's still worth sorting...

I think the example in the description is a valid use case and an example of why we can't just leave it to the implementation (because if an implementation doesn't...

Just a thought, what would you think of a list comprehension expression, eg something like the python syntax? (just throwing out some ideas here): ```wdl # 1: Array[String] arr_with_suffix =...

@ruchim if we go down this route, I don't see why not!

@patmagee @geoffjentry are we accepting future version PRs yet? Side question: is there a work-in-progress name for the next version to avoid proliferating not-quite-`version 1.1` WDLs (like we did with...

> Redesigning WDL to only use environment variables and deprecate the templating is the better option. That's a really interesting concept. It would let us remove the (command) interpolation syntax...

If your TSV input could instead be `json`, you can use `read_json` to get objects with headers. Another advantage of doing that: when `struct`s come in with draft-3 you can...

Oh I see. Theres no reason that this is a bad idea, I was just mentioning JSON in case it was useful in the meantime :)

* IMO the "least surprising" behavior would be that all whitespace in the multi-line string are preserved by default. * I like the idea of starting a line with `|`...