From: Matthew Boeh <mboeh@...>
Date: 2007-11-11T19:18:12+09:00
Subject: Re: method names in 1.9


--tNQTSEo8WG/FKZ8E
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Sun, Nov 11, 2007 at 06:09:21PM +0900, Trans wrote:
>=20
[snip]=20
> > - to_splat
> >
> > Unlike invoke_method above, you didn't provide your sought name.  It
> > must be a clear name to be use for splat.  Any opinion?
> > Regarding 'to_xxx', we have 'to_open' added in 1.9 that must be
> > irritate you.  It is used to implement open on anything other than
> > strings.
>=20
> I tend to agree with David here. For instance, I once had
> Integer#to_roman, but depraecated it in favor to Integer#roman. Why is
> the preposition needed? #splat seems fine. On the other hand one could
> just argue that a "splat" is a form of Array, so #to_splat is just an
> odd named part of the to_a and to_ary "family" of methods. Would
> #to_args make us feel any better?
>=20
> T.
>=20
>=20

I have a few thoughts here.

1. The splat, although I love it dearly, is a generally nonintuitive concep=
t. =20
Considering how ludicrous #to_list_of_arguments would be, it becomes necess=
ary=20
to choose some sort of memorable term for the concept. I think "splat" is s=
uch=20
a term -- the asterisk looks like something that's been splattered, the=20
process of separating the elements of an array into arguments of a function=
 is=20
consistent with the term "splatter," and it's an active-sounding and unique=
=20
term for an unusual operation.

2. That said, I can't think of a valid use case for implementing #to_splat=
=20
separately from #to_a that isn't tantamount to direct hostility against any=
one=20
who has the pleasure to read your code. Many Ruby programmers would be=20
genuinely confused by a use of the splat on something that isn't an array, =
and =20
the majority of the remainder would likely expect the semantics to be=20
identical to #to_a.

3. Finally, I have to admit to surprise for why the asterisk isn't consider=
ed=20
a unary operator and treated as such. That is:

  class Confusing
    def to_a
      [1,2,3]
    end
   =20
    def *@
    # that is, analogous to -@, +@
      to_a.reverse
    end
  end

If I were to unexpectedly come across code like this:

  func(*ary)

without any understanding of the splat, I'd probably look first for an=20
implementation of the unary asterisk as an operator. def *@ looks like line=
=20
noise, sure, but not significantly more so than -@, +@, etc.

--
Matthew Boeh

--tNQTSEo8WG/FKZ8E
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Digital signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQFHNtbnfi+i1R7d/ZwRAuqDAKCEQTmTDpyP9w0a6WdRaHmxSvncfwCgn7KH
bvbEYPgOKxuXnjdbfD5ZCRY=
=SL7y
-----END PGP SIGNATURE-----

--tNQTSEo8WG/FKZ8E--