It's kind of funny, because my feelings are the opposite of what blokhead describes in the last paragraph of Re^3: Representing all data as Lists; I'd actually be more inclined to support unifying the internals than changing the interface, though I don't really think either is a very good idea.

Like I said (or maybe I should have said more clearly) in my original reply, I think different things should look different. You may say the corollary is that alike things should like similar -- and I'd agree -- but in this case I dont' think a string and a list are things that are really all that alike. I can understand why someone would say they are, but I think it's a contrived likeness.

I strongly agree with Perl's current notion that a string is an individual thing (i.e. a scalar). I don't often have the need to break that thing apart and operate on its individual characters (Update: a clarification of what I meant). Perhaps that's a learned trait, but I tend to think it's more likely that instinctively seeing a string as a list is less natural.

In the case of hashes, I don't feel as strongly as I do about strings. Hashes are very list-like already, and as you pointed out, you can already coerce a list into a hash. That's a pretty natural-feeling thing, in my eyes. I still think a hash is different enough from a plain list that it deserves its own syntax and sigil, though. And Perl 6 will make it pretty easy to get a list of pairs, or a list of interwoven keys/values:

my %hash = <one 1 two 2 three 3>;

say $_.join(" ") for %hash.pairs;

# output:
#   one 1
#   three 3
#   two 2

say for %hash.kv;

# output:
#   one
#   1
#   three
#   3
#   two
#   2
[download]