Hey guys - here's a question on Zend Framework or better on MVC in general:
I am asking myself for a quiet a long time now, if it is a good idea to push business objects (User, Team, etc.) to my views or if it would be better just to push dump data containers such as arrays to the view for rendering.
When pushing business objects to my view I have a much tighter coupling between the views and my domain model, however, the view could easily do things like foreach($this->team->getUsers() as $user) { ... } which I personally find very handy.
Providing domain model data in dumb arrays to me looks more robust and flexbile but with the costs of that the view cannot operate on real objects and therefore cannot access related data using object's method.
How do you guys handle that?
Thanks much,
Michael
It's better to make your View access a Domain Model object in an object-oriented manner, instead of using the Controller to convert Model data into plain scalars and arrays.
This helps to keep the Controller from growing too fat. See the Anemic Domain Model anti-pattern. The Controller only needs to know what Model to instantiate, passes the request inputs to that Model, and then injects the Model into the View script and renders. Keep in mind that a Domain Model is not a data-access class.
You can also write View Helpers to encapsulate a generic rendering of a Domain Model object, so you can re-use it in multiple View scripts.
Your View should accesses the Domain Model only in a read-only manner. View scripts should not try to effect changes to the Domain Model.
You can also design your Domain Model to implement ArrayObject or other SPL type(s), as needed to make OO usage easy in the View script.
It's true, a large driving motivation of MVC and OO design in general is decoupling. We want to allow each layer to remain unchanged as the other layer(s) are modified. Only through their public APIs do the layers interact.
The ViewModel is one solution to abstract the Model so that the View doesn't need to change. The one I tend to use is Domain Model, which abstracts the details of table design, etc. and supplies an API that is more focused on the business rather than the data access. So if your underlying tables change, the View doesn't have to know about it.
I would expect that if there's a change to the Domain Model, for instance it needs to supply a new type of attribute, then it's likely that your View is changing anyway, to show that new attribute in the UI.
Which technique you choose to decouple one layer from the others depends on what types of changes you expect to be most frequent, and whether these changes will be truly independent changes, or if they will require changes to multiple layers anyway.
The "standard" approach would be to completely prepare the model in the controller (e.g. fetch all teams, including users) and then send that to the View for presentation, but you are not bound by that. The data structures can be whatever you want it to be: Array, ArrayObject or custom Classes - anything you deem appropriate.
I dont use Zend framework, so this is in repsonse to the general MVC Have a look at the ViewModel pattern.
http://www.lostechies.com/blogs/jimmy_bogard/archive/2009/06/29/how-we-do-mvc-view-models.aspx
I'm comming from a .Net MVC point of view but I believe the concepts will the same.
I will do all my view rendering in the controller bascially like below
model only output dataset/objects (this should contain the most code)
controller assign view and add necessary HTML and make use of models
view only contains placeholder and other presentation stuff and maybe ajax call
So my team can work on each part without interrupting each other, this also add some information security to the project i.e no one can retrieve all the working code they only communicate by variables/object spec.
Related
Is it a good practice to use Spring Controller class(with #ModelAttribute) and the jsp to prepare model at the same time or the model has to be prepared only by Spring and the view from the jsp?
The idea comes from this topic . I have a Controller class:
#RequestMapping(value = {"", "/"}, method = RequestMethod.GET, params = "mode=create")
public ModelAndView showCreatePage(#ModelAttribute("createForm") ApplicationCreateForm form)
{
return customMethod("some string");
}
and in my jsp I have:
<jsp:useBean id="createForm" scope="request" class="com.example.ApplicationCreateForm"/>
I do not need to populate the form with the information to be present to the user all fields are empty.
So from what I uderstand I have been declared ApplicationCreateForm bean twice, with the same scope - request.
Is it a good design practice to use both at the same time? Is there a reason for that? Does the second declaration(in jsp) give me more power, for example to override the model or it is complete unnecessary? Does the second declaration overrides the first one when both are present?
There are many things wrong with this implementation.
Is it MVC?
If JSP know about Model, why do we need controller. Lets remove the routing engine and use JSP directly to consume Model. But then the application will be monolithic. I believe you do not want that. We have two major flavours of MVC. In both, controller is the front facing object. It receives the command, interprets it, works with data layer and gets Model. If required the Model gets converted into a viewModel and then this object is passed to the view.
Why viewModel?
Say we are implementing paging on screen. We are showing list of persons. Person is your model here but your view also need to know page number, page size etc. Your model thus may not fit directly in this case.
Say we need data from multiple tables to shown on screen. This data is related in some way. Now will you pass separate model objects to view and let it do all the business logic? Ideally no.
Should not the design support DTO or ViewModel or Commands and Queries?
We want our application to be designed properly. As stated above we need to send data to view or clients (REST) after processing. Processed data may not map to you domain until unless we are just creating a CRUD stuff. What if you want to implement CQS or CQRS?
Where is separation, where is SOLID?
If my view is performing business logic then where is 'S'? If view knows about model and need to be changed in slightest of changes in model then where is 'O'?What if I want to separate queries from command (CQS) and scale the two things separately?
To conclude I would say, yes you can do this but it is not good if you are developing a decent size application or you think it will eventually be one. I have seen people using model entities starting from ORM, to controller to view. It will work but the question is do you want a monolithic and tightly coupled application. As per my experience the presentation logic (view) has very different logic and need of data in comparison of controller and same goes for your data access layer.
I think you should prepare your model fully in spring controller. Then view should be as passive as possible, ie. only showing model attributes received from controller while having no further knowledge about application logic. This approach gives you clean separation of concerns and you view is as independent as possible and can be easily switched for different view technology - eg. thymeleaf or freemarker templates. The whole idea of MVC is to separate presentation layer and by leaking business logic to view you create unnecessary dependencies.
Your view should be as simple as possible, as the logic leaked to view makes it very hard to test and reuse. On the other hand, if your logic is nicely separated, you can test it easily and reuse it easily. Ideally, business logic should be completly independent on web environment.
By defining you are creating tight coupling between your view and class com.example.ApplicationCreateForm, using spring mvc you achive loose coupling between your controllers, view and model it might happen that you change you model name but you still have same properties in it which might be enough for view, in above case you will need to update your view but it won't be required in case you are using Spring MVC.
Since spring comes with the concept of making things loosely coupled to make your code more testable, you should always keep in mind that separation of concern is your priority. So it is always the best practice to prepare your model fully in Controller, not in views.
Keep thing simple , make your controller responsible for preparing your model, and keep your views to display the model only.
So, a big NO to your question. The second declaration isn't going to make you powerful, rather help you to be tied up.
Wouldn't creating ViewModels lead to redundancy? In the sense I have my domain model and I need to display the data from it on a view. So we create ViewModels, add DataAnnotations to it and display it on the View. At this point I have 2 object with almost identical data.
As others have already said, only the most trivial application can get away with passing their domain models directly to the view. Even then, it's still not a good idea for a lot of reasons.
The requirements of your view are different from the requirements for your data model. For instance, you may have a field that is required in your view, but is nullable in your vie model. If you add a `[Required]' attribute, then your model will now consider this field non-nullable.
However, my single biggest reason for never using domain models in views is for security. MVC allows you to post any value to it, and the default model binder will happily plug values you post into the model, which means if you had an IsAdmin flag, and someone posted a true value for IsAdmin, then when you saved the changes to the model, someone is now an admin.
The first rule of web security is never trust input from the user, and passing your view models directly to the view basically gives up sanitizing your data.
Yes, it is a form of redundancy. But redundancy is often required to achieve other goals. In case of Models, having this separation of view models and domain models helps in achieving a decoupled setup between your view and data-store. And it is not often that ViewModels are exact copies of Domain.
Which means, either can change without having an impact on other. I can see cases where this would be valuable - data-type changes in table need not call for deployment of the web application.
So, in summary, while there is redundancy, it is a design choice on whether the system is complex enough to benefit from this redundancy.
In 99% cases ViewModels don't lead to redundency.
The only 1% which comes to my mind is simple application with anemic domain models and pages, which shows nothing but a single model on a page. This is peculiar to content management pages.
In any other case:
1) your ViewModels will join information from multiple domain models
2) you'll have a logic specific to your domain in domain models
3) it's not a good idea to mix view-specific metainformation like DataAnnotations into your domain
Nope, using ViewModel has its own purpose. Let's think about a situation when your view has two or more Entities from the Domain Model, without a ViewModel, how are you gonna organize data? The data needed for a view sometime is not exactly like domain model, it can has
less or more information and sometime we have to pre-process data from domain before rendering view(e.g. format date time depends on client's culture).
Furthermore, ViewModel help de-couple the Web UI from the domain layer. Entities in Domain Model is not just about data representation(which is the only purpose of view model), they also have operations that mimics the business rule, you don't want expose too much domain knowledge to the UI layer who doesn't need to know.
Could anyone give an example of why it would be advantageous to use MVC instead of a simpler Model and a View only.
Note: whether it's called MVC or MVP (Model-View-Presenter), I'm talking about the one where the View receives input, then the Controller will respond to the input event by interpreting the input into some action to be done by the Model. When the model changes, the View will update itself by responding to events from the model.
What is disadvantageous of simply letting the Model respond to events in the View and vice versa?
In MVC, if I changed the model in a way that affects the controller then I'll have to do changes in the controller. In Model-View, if I change the Model, I'll have to update the view.
So, it seems like we are introducing complexity by adding the "controller" part?
In MVC, the Model is blind to its environment, the view can be too - passing off (blindly) its events to the controller, which knows more about the view and model. So when all is said and done, the controller is the 'non-reusable' disposable part of the system, since it is the most context aware component.
if I changed the model in a way that affects the controller...
The the model should expose simple CRUD methods in such a way that those using the methods do not have to know anything about the passed update object, nor what really happens inside the model.
This means that the view, IMO, has to do a bit of work by creating the passed record, since Controllers are supposed to be stateless and the view is more persistent. Controllers get triggered and 'kick-in' do their work with a passed object and do not have a state.
The passed data is created by some sort of generic convention.
Let me go even further. Suppose you have a view, a tablegrid, and a control whose enabled property is dependent on item is selected in the grid -- you COULD create a view that handles both those controls and this logic internally, and that would probably be the way to go in such a simplified example.
But the more atomic your views are, the more reusable they become, so you create a view for every, yes every, control. Now you are looking at a situation where views have to know about each other in order to register themselves for the right notification...
This is where the controller steps in, since we want to stick all these dependencies onto him, the long term disposable one. So the controller manages this type of view-to-view notification scheme.
Now your views are ignorant as they can be and independent, thus reusable.
You can code a view without having to know about the system, or the 'business logic' as they like to call it. You can code a model without having to know too much about your goals (though it does help to tweak the model to enable it to return the datasets you have in mind).... but controllers, they are last and you have to have the previous two firmed up before you can wire things together.
Here is another thing to think about -- just as the Model is supposed to abstract-away and provide a generic interface to the underlying implementation of the data it is managing (the client does not know if the data comes from a DB, a file, a program setting, etc) -- the view should also abstract away the control it is using.
So, ultimately this means a view should not (caveat below) have functions/properties that look like this:
public property BackgroundColor{get;set}
Nor
public function ScrollBy(x,y){}
But instead:
public SetProp(string name, object val){}
And
public DoCmd(string name, object val){}
This is a bit contrived, and remember I said ultimately... and you ask why is this a good idea?
With reusability in mind, consider that you may one day want to port things from WinForms to, say, Flex, or simple want to use a new-fangled control library that may not expose the same abilities.
I say 'port' here, but that is really not the goal, we are not concerned with porting THIS particular app, but having the underlying MVC elements generic enough to be carried across to a new flavor -- internally, leaving a consistent and ability-independent external interface intact.
If you didn't do this, then when your new flavor comes along, all your hard references to view properties in the (potentially reusable/refactorable/extendable) controllers have to be mucked with.
This is not to mean that such generic setters and cmds have to be the interface for all your views abilities, but rather they should handle 'edge case' properties as well as the normal props/cmds you can expose in the traditional hard-link way. Think of it as an 'extended properties' handler.
That way, (contrived again), suppose you are building on a framework where your buttons no longer have buttonIcon property. Thats cool because you had the foresight to create a button view interface where buttonIcon is an extended property, and inside the view your conditional code does a no-op now when it receives the set/get.
In summary, I am trying to say that the coding goals of MVC should be to give the Model and View generic interfaces to their underlying components, so when you are coding a Controller you don't have to think to hard about who you are controlling. And while the Controllers are being (seemingly unfairly) set up to be the sacrificial lamb in the long run of re-usability -- this does not mean ALL your controllers are destined for death.
They are hopefully small, since a lot of their 'thinking' has been shoved off into semi-intelligent Models and Views and other controllers (ex: Controller to Sort a Grid or Manipulate a TreeView) -- so being small they can be easily looked at and qualified for reuse in your next project -- or cloned and tweaked to become suitable.
It actually reduces complexity by separating the workflow logic from the domain logic. It also makes it easier to write unit tests and makes your application easier to maintain and extend.
Imagine if you wanted to add a new data type. With the approach above, you would probably duplicate a lot of the workflow logic in the new class as it would be likely to be tightly coupled to the domain logic.
The discipline involved in separating the workflow logic into the controller makes it more likely that you will have fewer dependencies between workflow and domain logic. Adding a new data type would then be more simple, you create the new domain object and see how much of the controller you can reuse, e.g. by inherited from a controller super class.
It would also make it easier to change frameworks in future - the model would probably not change too much and so would be more portable.
Having said that, you might want to look into MVVM depending on what you are using as your presentation layer: Benefits of MVVM over MVC
Advantages of MVC/P (I am talking about Supervising Controller here) over MV include:
You can handle complex data binding code in the controller, if required.
You can test that complex presentation logic without a UI testing framework.
You can also have a graphic designer make your views, and not see your code, and not mess up your code when they fix your views.
I've read some MVC advice in the past regarding models stating that you should not reuse the same model objects for the domain and the view; but I haven't been able to find anyone willing to discuss why this is bad.
It is my opinion that creating two separate models - one for the domain, one for the view - and then mapping between them creates a lot of duplication, plus tedious mapping code (some of which might be alleviated by things like AutoMapper) that will likely be error prone.
What makes having a separate model for the two concerns worth the trouble of duplication and mapping code?
At its heart, two models is about Separation of Concerns. I want my View to work off of a single Model. I want my Domain Model to represent the conceptual model I build with the domain experts. ViewModel often has technical constraints. Domain Model is about POCO, and not being bound by technical constraints of either data shown (View) or persisted (in a DB or otherwise).
Suppose I have three entities shown on a screen. Does that mean I need to force a relationship between the three? Or just create a ViewModel component object that contains all three items. With a separate ViewModel, View concerns are separated from my domain.
why? 'cause the view shouldn't have the ability to use the model object!
Imagine you pass the project to a web designer to do the view layer. Suddenly he/she has the ability to mess around with your application's data through the model layer. That's not good.
So always only pass the data the view needs, instead of the object with methods.
J.P. Boodhoo's article Screen Bound DTOs will help you understand the design benefit.
There is also a security benefit that I have written about.
Having a presentation model simplifies your views. This is especially important because views are typically very hard to test. By having a presentation model you move a lot of work out of the view and into the domain->presentation model. Things like formatting, handling null values and flattening object graphs.
I agree that the extra mapping is a pain but I think you probably need to try both approaches in your specific context to see what works best for you.
There are even plainer concerns including the view model's ability to be specially formatted and certainly null-safe.
I guess the idea is that your domain models might extend to other implementations, not just your MVC application and that would break the seperations of concerns principle. If your View Model IS your domain model then your domain model has two reasons to change: a domain change requirement AND a view requirement change.
Seems I have duplication of rules as well.
ie. client object validation on the UI, then mapping to domain object that has to be validated.
What I tend to do however is map my set of domain objects to create a model - ie. a webpage that shows customer info, stock info, etc... my model becomes a structure that holds a Customer object and a Stock object.
CompanyPageModel
public Customer Customer{get;}
public Stock Stock {get;}
then in my mvc project ViewData.Model.Customer.Name ViewData.Model.Stock.CurrentStocks
Separation 'seems' like more work, but later, it's good to have this division of UI/Domain model...sorta like writing tests :)
I have drunk the cool-aide finally, I do like being able to mark my viewmodel with display instructions and have that all auto wired up.
What I demand now is some kind of auto generator of viewmodels from poco entities. I always set some string as an int and it takes me forever to find it. Don't even think of doing this without automapper unless you like pain.
I've done a fair amount of work on MVC on the web, and we're learning about it in my OOP class. I'm seeing some differences, and I can't tell whether that's because the Web's version of the MVC pattern is different than the traditional one, or whether I misunderstood it.
From my understanding, The model (your flat files, RDBMS', etc) is a generic data-housing object. The View (Browser, HTML, etc) is what the user interacts with, and the controller mediates between the users actions and the data. The controller is the most domain-specific part, and it manages the views, tells the model what it needs, and tells the views what to display.
In class, we have the Views matching what I just described, the Model contains a list of the views so that it can update them when the data changes, and the controller simply maps the user's actions to calls to the model and to specific objects (which may themselves, ask the model to update the views). What ends up happening is that most of the business logic is in the model, and it's tied very heavily to the simulation or application that is being written, while the Controller is reduced to a mapping tool between commands and methods.
What are your thoughts on this?
In a non-web interface the controller handles the inputs from things like the keyboard and mouse, choosing which views to render and what changes to make in the model based on those inputs. The view and model can be more closely related because the view can register callbacks directly with the model entities to be notified of changes and thus be updated based on changes to the model directly instead of being updated by the controller.
In the web world, views are necessarily more decoupled from the model. It must act through the controller actions because it has no direct access (after being rendered and delivered to the browser) to the model. The controller takes a larger role in this environment even though the only "input" it has to deal with are browser requests. In a sense, the coupling that used to occur with the view in non-web MVC is transferred to the controller acting on its behalf. Since there are no callbacks from the model to respond to (let's forget about "push" technologies for now), more business code is incorporated into the controller since it's the natural place to model business processes, though perhaps not validation.
In my understanding controllers in the web MVC pattern are just bridges between Models and Views, they simply grab the data from the Model and pass it on to the View. The Model and the View and independent and never talk to each other.