If push @a; were currently a syntax error, then it would make sense for it to mean push @a, $_; because doing something useful is better than a compilation error. But push @a; isn't a syntax error — it actually does do something — it pushes the empty list onto the array.

Pushing an empty list might seem like a no-op which could never be useful, but if you think that, you're forgetting about tied arrays.

Why isn't push @a; equivalent to push @a,$_;

Likely because that could cause ambiguity whether my @x=(); push @y, @x; should push $_ or not.

Sure, in theory it's possible for the compiler to differentiate the two. But I doubt that'll get changed in Perl 5, where the focus is on backwards-compatibility.

randomly,it seems that sometimes good ole $_ works

When $_ works is generally well-documented for each function.

Actually that was one of my first questions here.

(EDIT found it: why does push not default to $_?)

My guess is that it's because it's taking multiple args via push @a, LIST , and not a single one push @a, EXPR °

> allowing code to be more concise, succinct, and economical.

Nah, that's not true, because the defaults in Perl are not orthogonal.

You have to check for every operator and built-in how it's implemented.

especially the scalar behavior is very diverse.

EDIT

°) While it's still technically possible, this might be logically to confusing.

And symmetry would require a void ;shift or ;pop to automatically set $_ , which would most certainly break things!

push @_,$_;
[download]

Rather than judging what push @a; does against what you consider the proper implicit behavior, I think it's more appropriate to judge against what unshift does in this case.