Isn't this off-topic in SML-DEV? At any rate...
Clark C . Evans wrote:
> | IIRC, relative (.foo) types were very problematic.
> | <Oren> I'm beginning to wish I never brought them up.
> Hmm. I think they were problematic since we were
> trying to be too smart. The 90% use case is to
> name a different "leaf class" when you are
> within a well known island.
We did get into some pretty dark corners...
> | So we eliminated them for YAML 1.0, which IMO was a great idea.
> IMHO, the .foo (simple relative) variant is
> just too useful to drop....
> | KISS, Brian
> TMTOWTDI, Clark
Now that's a nice role reversal :-)
I figure we've already made a strategic decision to be "human friendly" at
some cost to "simplicity of syntax", so we can't immediatly rule it out at
this point. The change in the spec is minimal - some wording in the transfer
method property section, plus a single new production.
So, I'm willing to go with it if you manage to convince Brian :-)