This is in the continuation of Small classes and methods, but code still difficult to maintain and follow as I have matured the problem and will rebound on people's answers.

Here are all the elements that should be modeled:

1.Model (in this case a link directory)

2.Validations (of the model fields from user submitted form)

3.Database CRUD operations of the model

I'm thinking of properly organizing all these elements into classes to respect the traditional principles of abstraction, loose coupling and encapsulation.

Here is how I did it currently, which is uncomfortable to maintain because of a loose coupling problem or so I think: when I must perform an operation, I see myself jumping around:

  • Module 1:
    • Link model class: with only data members
    • Validation function that takes in all models data, and return either a bool if no error are found or a message back with all validation error. It does not contain the validation logic, which is contained in another module 4.
  • Module 2: the controller file It grabs the CRUD http operations and send instructions to a utility service file that holds the database code that will perform the CRUD operations.
  • Module 3: the database utility service file This takes the instructions sent by the model and perform the crud operations. It is un the form of a simple function.
  • Module 4: the validation logic service The validation code in the model just send the data from the model to the validation logic service file here.

From the answers of gbjbaanb in the other thread, I think my problem is a coupling and encapsulation problem. he said:

you should be thinking in terms of making the entity self-contained, so you can think of making the entity and operating on it in much simpler terms, like creating it is one action, validating it another and so on

Which I think is a very important point. Thus I plan to remodel my structure as such:

Module 1:

  • Link model class
    • data members
    • Create method
    • Update Method
    • Read Method
    • Delete method

Module 2: the controller This will simply grab the http requests and instantiate corresponding model class method.

A question I have is, how to handle validation ? Logically I would put code in the model class as a method, since its coupled with model. But validation code is quite complex and is in own service file, so how to organize it, considering that validation could be reused by other classes ? The current way with validation function in model module, and logic in service file, is very unpractical.

As for the database code, I think I should leave it in it's own module, which would create a coupling between model and database class, but I think it's unavoidable as it will be used by other classes.

Also, if you have another suggestion to organise this, please advise. For example, would you put the crud operations in the model as I did, or in the database class ?

1 Answer 1


Validation is a pretty tricky thing to handle but I've seen most frameworks handle it in a similar way. In the frameworks I've worked with, data validation works by annotating your model classes. This 'technique' of validation is good because it follows DRY or "Don't Repeat Yourself".

The best form of suggestion is through example:


Data validation for a model is added by adding data annotations to the model classes:

Example taken from here:

public class Movie {
    public int ID { get; set; }

    [Required] // <= the data annotation!
    public string Title { get; set; }

    public DateTime ReleaseDate { get; set; }

    public string Genre { get; set; }

    [Range(1, 100)]
    public decimal Price { get; set; }

    public string Rating { get; set; }

Then the a controller validates it like this:

// POST: /Movies/Create

public ActionResult Create(Movie movie)
    if (ModelState.IsValid)
        return RedirectToAction("Index");

    return View(movie);

The data validation modules are part of the framework and I personally have just used them without really thinking about their implementation but I'm sure you can get an idea of how to create your validation modules from their docs. Here is some documentation on C# Data Annotations.

Java EE

I'm not too familiar with the Java bean validation spec (JSR-303 and JSR-349) but I know that it also handles validation similarly.

Example taken from here:

public class User {

    private String email;

    @NotNull @Email
    public String getEmail() {
        return email;

    public void setEmail(String email) {
        this.email = email;

public class UserService {

    public void createUser(@Email String email,
                            @NotNull String name) {

These are just some examples of how some frameworks handle validation. The thing to keep in mind is that you should keep the concept of DRY - "Don't Repeat Yourself" and as beanvalidation.org states it

Constrain once, validate everywhere

If you come up with validation that works in this way, you're doing it correctly :D I hope I've lead you in the right direction. Learn from example!

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.