Comments on: Day 6 – The X and Z metaoperators

By: colomon

colomon — Thu, 09 Dec 2010 19:00:50 +0000

Ack, I must have been mostly asleep not to notice that this morning. :)

By: timtoady

timtoady — Thu, 09 Dec 2010 18:56:22 +0000

As for why it says 2 4 2 4 rather than 1 3 2 4, that’s just a rakudobug that appears to involve accidental aliasing of returned variables.

By: colomon

colomon — Thu, 09 Dec 2010 14:29:58 +0000

Because (1, 2) Z, (3, 4) is the equivalent of ((1, 2) Z, (3, 4)).Str, which is doing something along the lines of "Take each element in the list, convert it to a string, and return them separated by spaces." .perl is more akin to Data::Dumper.

By: prakash

prakash — Thu, 09 Dec 2010 14:18:46 +0000

$ perl6 -v

This is Rakudo Perl 6, version 2010.11-15-gfedc117 built on parrot 2.10.1 RELEASE_2_10_1-679-g9bec614

$ perl6
> ((1, 2) Z, (3, 4)).perl
((1, 3), (2, 4))
> (1, 2) Z, (3, 4)
2 4 2 4

Why are the outputs different?

Thanks.

By: Larry Wall

Larry Wall — Mon, 06 Dec 2010 23:26:01 +0000

Some additional differences are that hypers can work dwimmily on hierarchical values, while X and Z cannot, because they’re more list oriented. Also, since X and Z are looser than comma, you don’t have to parenthesize comma lists as you must with (most) hyper infixes. That is, hypers are transparent to the precedence of their base operator, while X and Z (and metasequences based on them) are forced to list infix precedence.

By: Matthew Walton

Matthew Walton — Mon, 06 Dec 2010 21:42:56 +0000

I certainly can. It’s just about output formatting for the sake of these examples. The .perl method in Perl 6 returns a string representation of the object which could be fed back through the compiler to reproduce the same object (in theory – it doesn’t work for absolutely everything. It’s a bit like Data::Dumper, if you know that from Perl 5). So I used .perl when I was writing the post in order to get the output in a form which shows the structure of the produced lists. If I’d omitted it, I’d have got the standard results from converting a list to a string, which would be something more like:

> say (1, 2 X 3, 4) 13142324

Far less useful for the illustrative examples in the post.

By: pete_mcstubbins

pete_mcstubbins — Mon, 06 Dec 2010 20:19:16 +0000

Can anyone explain why the “.perl” is needed at the end of an expression for the code which begins with “say”? Thank you in advance.

By: colomon

colomon — Mon, 06 Dec 2010 15:39:37 +0000

No, it’s only “equivalent” in the simple cases. The big difference is that zip is lazy and sequential, while hyper is eager and may execute in any order.

By: Dean

Dean — Mon, 06 Dec 2010 15:22:09 +0000

Is the Zip operator always equivalent to the corresponding hyper-operator?

> (1,2) Z (3,4)
1 3 2 4
> (1,2) >>,< (1,2) Z+ (3,4)
4 6
> (1,2) >>+<< (3,4)
4 6

By: Christoph

Christoph — Mon, 06 Dec 2010 10:37:54 +0000

> X can only handle an infinite list on the left, otherwise it would never manage to get anywhere at all.

This restriction only applies because S03 demands ordered results (“The returned lists are ordered such that the rightmost elements vary most rapidly.”) If you drop this requirement, X could easily take infinite lists on both sides (cross-products of countable sets are still countable)