Forum: Ruby on Rails Accessing Models via Views

Announcement (2017-05-07): www.ruby-forum.com is now read-only since I unfortunately do not have the time to support and maintain the forum any more. Please see rubyonrails.org/community and ruby-lang.org/en/community for other Rails- und Ruby-related community platforms.
(Guest)
on 2009-04-20 00:35
(Received via mailing list)
Hi

When creating a view to display infromation obtained from a model,
should the view always get this information from the controller, or
should it get it directly from the model?

For example, say I have Class and Student models, and they are in a
many to many relationship via the enrolled table. Now my view wants to
display the students of a given class. Should the controller run the
query and store it in a variable (say @students) that is passed to the
view? Or, should the view just access it directly like

<% for enrolled in @class.enrolled_students %>
<%= enrolled.student.name %>

I like the second way because it does not require me to put any code
in the controller. I just feel like this is coupling the view and to
the model and I can't tell if that is a bad thing in an MVC framework?
I feel like it is.

Thanks in advance
Phlip (Guest)
on 2009-04-20 01:39
(Received via mailing list)
removed_email_address@domain.invalid wrote:

> When creating a view to display infromation obtained from a model,
> should the view always get this information from the controller, or
> should it get it directly from the model?

Which one is simplest?

I think it's ideal when the controller sets up a single variable, such
as
@datagram, and then the view uses only first-level method calls on
@datagram:

   @datagram.marbles.each do |marble|
     <%=h marble.name %>
   end

Simplified eRB, of course, but you get the idea. We did not call
@datagram.method.method(42).method.marbles. The model and controller
should have
set any parameters up for @datagram to do its job, such as provide the
42.

(.each is technically a method call, but it's more stable, and it's
essentially
a part of Ruby's core mechanics, so we give it a pass. We would not give
.select{} or .inject{} the same pass!)

> For example, say I have Class and Student models, and they are in a
> many to many relationship via the enrolled table. Now my view wants to
> display the students of a given class. Should the controller run the
> query and store it in a variable (say @students) that is passed to the
> view? Or, should the view just access it directly like
>
> <% for enrolled in @class.enrolled_students %>
> <%= enrolled.student.name %>

I can think of no reason to use 'for in' in Ruby. I suspect Matz added
it early,
  to imitate languages that must use 'for'...

But your enrolled.student.name (is missing an h, and) has too many dots.
I
suspect that would violate something called the "Law of Demeter" if any
of those
dots took a "left-turn" into another high-level module.

Because you will find enrolled delegating to student frequently, your
code is
already more dry if enrolled has a def name; student.name; end. That
hides a dot
in the model, which is the ultimate goal of refactoring - dense models,
aggressive but sparse controller actions, and thick views with minimal
Ruby
method calls.

The controller is like the Government. Paraphrasing Abe Lincoln, it
should only
do what the model and view can't do for themselves.

> I like the second way because it does not require me to put any code
> in the controller. I just feel like this is coupling the view and to
> the model and I can't tell if that is a bad thing in an MVC framework?
> I feel like it is.

Coupling the view to the model is part of MVC. Look up one of the
diagrams - I
think there's an arrow from the V to M...

--
   Phlip
   http://flea.sourceforge.net/resume.html
(Guest)
on 2009-04-20 03:34
(Received via mailing list)
Thanks Philip. Your comments have helped a lot.

I totally forgot about the Law of Demeter learned way back in second
year CS. If I just follow that rule then I think most of my complaints
about doing what I was in the view will disappear because I will only
ever be using 1 level of indirection.

I will have to consult the pick-axe again on the nuances of iterators.
I haven't been using them because coming from a major C/Perl style
background, for is just more familiar to me. Thanks for pointing this
out though, no one else has yet.

As for the model <=> view in MVC, I was looking at 'Design Patterns'
and their is indeed a dotted line, where all the rest are solid lines.
The book fails to distinguish the difference between the two though.
(Guest)
on 2009-04-20 20:38
(Received via mailing list)
Thanks as well Rob. I like how you guys gave basically the same
response at the exact same time. It really shows a lot of uniformity
in the community. I was honestly expecting to get a lot of flame back
about 'do whatever is best' or something to that effect.

On Apr 19, 5:37 pm, Rob B. <removed_email_address@domain.invalid>
Ar C. (Guest)
on 2009-04-21 02:01
I tend to prefer code like:

@class.enrolled_students.each do |student|
   student.name
end

as Philip mentioned.

Enrolled students, if I understand your model(s) correctly, are just a
specialized set of students. Not dropped_students, or
withdrawn_students, or waitlisted_students, or students_on_hiatus, but
those currently enrolled in the class.

The ".each do |call it whatever|" gives you the chance to map the
'enrolled_students' back to their source, and pretty much anyone reading
the code will make the leap (however small) that 'name' is an attribute
or method of the 'student' class.
Phlip (Guest)
on 2009-04-21 02:08
(Received via mailing list)
Ar Chron wrote:

> I tend to prefer code like:
>
> @class.enrolled_students.each do |student|
>    student.name
> end
>
> as Phlip mentioned.

Tx - it reminded me to write up "Contractive Delegation" here:

   http://broadcast.oreilly.com/2009/04/contractive-d...

I heard the term from an old Smalltalker, and it stuck...

--
   Phlip
   http://flea.sourceforge.net/resume.html
This topic is locked and can not be replied to.