In generic item context an array returns a capture (think reference) to an array.

But in other item contexts, like boolean ?@array, numeric +@array and string ~@array the array will return different things (for example number of elements, and a string representation).

A scalar value that holds an array capture automatically dereferences, so you can write $ary[0] instead of the cludgy $ary->[0].

I don't know why this should make sigils useless. Sigils serve two purposes: visual distinction, and name disambiguation (%foo is a different variable than @foo).

Automatic dereferencing doesn't impact any of those.

It's like in Perl 5 - you can store everything in a scalar (or scalar ref), but you shouldn't, because it makes your program less readable.

But if you decide to do that in Perl 6, the syntax is nicer ;-)

Update: as a side note the braces after a variable are nothing special or magical in Perl 6, it just calls  method postcircumfix:<[ ]>($invocant:, $argument) on the variable. So it's just natural that it works on a capture as well as on an array.

Is the reason that scalar context has been renamed item context to give it some symmetry with arrays/lists? So now we can have arrays associated with groupings (list context) and scalars associated with single objects (item context)? I was futzing around the synopses trying to answer this, and didn't come to a conclusion.

Looking through all of this, I'm sort of shocked at how much Perl 6 I've missed out on. These are very interesting reading, and I wasn't aware that design was this far along. (I've not followed Perl 6 at all)

Sigils serve two purposes: visual distinction, and name disambiguation (%foo is a different variable than @foo).

Automatic dereferencing doesn't impact any of those.

I'm not convinced. What if an Object is a collection class of some kind? And it only is useful on the top-level, as anything already inside another collection is a scalar.

—John
Writing from Pudong Airport

What if an Object is a collection class of some kind?

Could you elaborate on why this is could be a problem?

Most people have been happy to drop the indirection operator -> between elements of a nested structure:

my @a = ( [ 1,2 ], [3,4] );
print $a[1]->[0]; ## => 3
print $a[1][0]; ## same thing
[download]

This just extends that convenience to the first level of dereference.

You have to jump through hoops, (assigning the ref to a glob ref), to achieve the same thing in Perl 5.

Funny, I started using '->' more as my Perl skills increased. Always seemed clearer to my eye.

But more to the point, this confuses me:

my @x = (1,2,3);
my $x = ['x', 'y', 'z'];
[download]


Pre-Perl 6 treats @x and $x as completely different variables. So I know that $x[1] is going to contain '2' and $x->[1] is going to contain 'y'. But according to the posts here, Perl 6 treats $x[1] as equivalent to $x->[1]. Just how is that supposed to work?

As best I understand it, because in Perl 6, arrays always use the '@' sigil.

So elements of @x are accessed using @x[1]

And elements of the array pointed at by $x can be accessed using $x[1].

The abiguity that is confusing you is because Perl 5 used $x[1] to access elements of array @x.

@x = ( 1, 2, 3);;
print @x[1];;
Scalar value @x[1] better written as $x[1] at
2
[download]

The explanation for that was that the use of the '$' signified that the element accessed was a scalar. But elements of arrays are always scalars, which makes the purpose of changing the sigil redundant. So in Perl 6 you don't.

In the same way, if it hasn't changed since I last looked, hash elements are addressed using %x{'foo'} (or %x<foo>) and not $x{'foo'} as now.

---

Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.

"Science is about questioning the status quo. Questioning authority".

In the absence of evidence, opinion is indistinguishable from prejudice.

"Too many [] have been sedated by an oppressive environment of political correctness and risk aversion."

If you access an element of @x, you simple write @x[0] - not ambiguous in any way.

It takes a while to get used to, but sigils are invariant in Perl 6.

No! http://en.wikipedia.org/wiki/Shanghai

Just a typo, sorry! That's what I get for trying to type Chinese without my wife's help.

Sorry I missed you in 昆明. I was there for an extended stay.

—John
back home in Texas