Re: syntax request

Hi --

On Mon, 7 Apr 2008, Rick DeNatale wrote:

> On Mon, Apr 7, 2008 at 3:47 AM, David A. Black <[e-mail]> wrote:
>>  and I have to say that this, in 1.9:
>>
>>   puts object
>>   .methods
>>
>>  has left me rather puzzled. If that works, then this should work:
>>
>>   ruby -e ".methods"
>>
>>  but it doesn't --
>
> Why would you expect this to work?  if you omitted the first line from
> the ruby program above and just had this one line, it wouldn't work
> either.

That's basically the example I've given. I don't know; honestly I
can't really puzzle it through, because it makes no sense to me. I
suppose I would expect it to call 'methods' on main or something. Or
maybe on the last value encountered in the last Ruby program you'd run
:-) This won't withstand close scrutiny; I know that no one said it
pertains to 'self'. Basically I just think it's an inconsistent
parsing rule that leads to an ugly result, so I can't get too enthused
about it.

>> which means that the whole thing introduces a new
>>  kind of "default object", other than self: an object to which an
>>  otherwise uninterpretable dot is supposed to connect.
>
> I don't think so. I think it means that the 1.9 parser has been
> altered to look past new lines looking for a dot as the first
> non-whitespace character.  It's purely a parsing issue.

I guess it's all parsing, in a sense. I see it as the parser being
altered to accomodate the idea of a kind of lazy evaluation: "the"
object (the last thing on the line) sort of hangs around waiting until
we see what's happening next, and how that object is treated *depends*
on what's happening next. That's what I don't like.

> That said, I'm not enamored of this new 'fluent' syntax.  The
> motivating use case seems to be to string a bunch of method calls
> together so that instead of

I'm not sure where the word "fluent" to apply to this came from
(though I've heard it before). It doesn't flow any more than any other
syntax -- a trifle less, I fear. And, as you say, it's in the service
of providing new syntax for something that isn't a great practice in
the first place.

>   a.b.c.d.e.f
>
> you can write:
>
>   a
>    .b
>    .c
>    .d
>    .e
>    .f
>
> Which to my mind reeks with a strong "smell of demeter" (I personally
> don't consider the "law of demeter" a law but rather a code smell to
> be investigated when it's violated).

I like Martin Fowler's statement that he'd like the Law of Demeter
better if it were called The Suggestion of Demeter :-)


David

-- 
Rails training from David A. Black and Ruby Power and Light:
   ADVANCING WITH RAILS   April 14-17          New York City
   INTRO TO RAILS         June 9-12            Berlin
   ADVANCING WITH RAILS   June 16-19           Berlin
See http://www.rubypal.com for details and updates!