From: Daniel Luz <dev@...>
Date: 2008-12-27T13:55:03+09:00
Subject: [ruby-core:20910] Re: named parameters?

I remember reading Matz's slides from 2005 about that:

    http://www.rubyist.net/~matz/slides/rc2005/mgp00010.html
    http://www.rubyist.net/~matz/slides/rc2005/mgp00011.html
    http://www.rubyist.net/~matz/slides/rc2005/mgp00012.html
    http://www.rubyist.net/~matz/slides/rc2005/mgp00013.html
    http://www.rubyist.net/~matz/slides/rc2005/mgp00014.html

I like his syntax proposition, though I'd prefer something closer to
Python's semantics: *args and **kwargs shouldn't repeat information
captured elsewhere, i.e., IMHO the "Keyword Arguments (3)" slide
should result in:

    baz()           # rest=[],a=4,b=0,keys={}
    baz(1)          # rest=[1],a=4,b=0,keys={}
    baz(a:1)        # rest=[],a=1,b=0,keys={}
    baz(a:1,b:2)    # rest=[],a=1,b=2,keys={}
    baz(1,2,b:2)    # rest=[1,2],a=4,b=2,keys={}
    baz(c:2)        # rest=[],a=4,b=0,keys={c:2}

But of course, arguing over them is pointless if they aren't currently
in question :) According to the slides, they were originally planned
for 1.9. Since they're a pretty big change (see next paragraph), I
guess it might be better now to wait for 2.0 than introduce them in a
minor release.

Named parameters (at least in this form) also have the downside of
clashing with implicit Hashes, making the two almost completely
incompatible. But I believe a script could be built to rewrite files
to turn implicit Hashes into explicit ones, making old code still work
(except, of course, for comments, autogenerated code and eval'd
strings). In the end, given the huge potential in code clarity and
performance gains, I think it's worth the migration pain.

Just my two cents. I, too, hope they're added someday.

--
Daniel

On Fri, Dec 26, 2008 at 23:28, Roger Pack <rogerpack2005@gmail.com> wrote:
> Just a curiosity. Are named parameters on the list for anytime?  That
> would be sweet.
> m/b something like
> method :a: => value
> or something
> Thanks!
> -=r