From: Brian I. <in...@tt...> - 2002-09-10 20:42:59
|
On 10/09/02 20:13 +0000, Clark C. Evans wrote: > In the current specification, YAML is Functional, > Typed, and Graph. Since YAML is a serialization > format and flattened, the other combinations may > emerge. Thus, one could strip the types, the anchors, > and forget about the equality constraints and end > up with a PairList Tree Untyped. And this is good. > > The question is... which combination of these are > we going to spend the bulk of our effort in the > future (beyond the intial YAML specification). +1 The question is not what we *can* do. We've barely scratched the surface. It's: "how can we write a specification that allows people to get going on implementations, without fear of doing it wrong." We need to get past the idea that we are going to describe up front, everything that people can and can't do with YAML. That ain't gonna happen. We just want to describe (and I vote as short/sweet as possible) how YAML should be thought to flow though the programs that process it. Cheers, Brian |