On Wed, Aug 10, 2005 at 11:24:02AM -0700, Dave Whipp wrote: : Do Sets get a sigil?
Probably not. <cough>unicode<cough> : I'd guess that % would be appropriate, because a : hash is simply "Set of Pair" where the membership equivalence class is : simply $^member.key. What syntax is used to associate the equiv-class : with a set? At our meeting there was much discussion of the relationship of hashes and sets. The resolution seems to be that sets are immutable values, and you still want to use a hash with a fixed true value to get mutable sets. It should be easy to interconvert sets and hash keys though. But anyway, if we assume sets are values rather than containers, they don't really need a sigil. (The problem with treating sets as containers is that you can't do set theory on them. When you add to a set, it is in fact a *different* set, and should have a different identity.) Larry
