S.O.L.I.D Design Principles Explained : Part 1

First things first. SOLID is an acronym of acronyms. Which means that S is an acronym for SRP and SRP is acronym for Single Responsibility Principle and so for the other 4 letters.

SOLID stands for

  • S - Acronym for SRP.
  • O - Acronym for OCP.
  • L - Acronym for LSP.
  • I - Acronym for ISP.
  • D - Acronym for DIP.

Note: This article is divided into 5 peices, one for each principle. The code for each article will be available and all the 5 articles work on the same code base. But at the end of each article I have taken a backup of the code base for clarity and you should be able to download the code.

This part explains the acronyms for SOLID and a brief introduction to Cohesion and Coupling is provided for a better understanding of SOLID principles.

Now for a more detailed description of what the preceding acronyms stand for:

  • SRP - Single Responsibility Principle: a class should be designed only to support one purpose or responsibility.
  • OCP - Open/Closed Principle: "Software entities should be open for extensions, closed for Modifications".
  • LSP - Liskov substitution Principle: "objects in a program should be replaceable with instances of their subtypes without altering the correctness of that program".
  • ISP - Interface Segregation Principle: "many client-specific interfaces are better than one general-purpose interface."
  • DIP - Dependency Inversion Principle: one should "Depend upon Abstractions. Do not depend upon concretions."

Any project that follows OOPs have to identify and design the classes and interfaces first and then define the methods and write the code to accomplish various functionalities. But developers often prefer to show their programming skills first and then think of design principles. That is so common in many projects and they end up writing redundant code for the same functionality or design classes in such a way that by the end of the project the addition of new functionality without disturbing existing code becomes highly impossible. That is because we did not spend at least a few hours on analyzing what we are coding.

For the preceding principles to be clearer we need to understand two important terms. Any software guy knows two words for sure: Cohesion and Coupling. I learned them first time in my bachelors when studying the Software Engineering subject and these two words convey a significant amount of valuable information to any developer in the software field.

Coupling and Cohesion from Software Engineering

Coupling refers to the degree of interaction among modules i.e. their inter-relationships.

Cohesion refers to the degree of interaction among statements in a module.

Where module in the preceding refers to a set of one or more contiguous program statements having a name by which other parts of the system invoke it and preferably having its own distinct set of variable names.

Good modules exhibits low coupling and high cohesion.

Understanding in layman terms

Let us assume you are planning to go a movie. Initially you thought you would go alone and later felt like going with a friend. So, you bought two tickets and went to the movie. Both the tickets are with you. Unfortunately your friend did not turn up in time and if you go inside to watch the movie, he will be not be able to enter the theatre without his ticket. So, here your friend is dependent on you for the tickets. Had you given his ticket to him and went to the movie theatre with your ticket, you would have been able to watch the movie by yourself and he could then join you even if he comes late and this would have been easy to do rather than you missing some part of the movie until he turned up. This is what coupling means.

Keeping tickets only with you is high coupling and it is difficult to accomplish your functionality (watching the movie from the beginning) without your friend. So, we have to minimize the dependencies among entities though they both communicate with each other or help each other in accomplishing certain tasks.

Now for Cohesion. Assume you planned to go to the movie by taking public transport. There is a direct bus at 4:00 PM and it drops you off at the theatre at 5:00 PM and your movie is at 6:00 PM. This would make your life easy without worrying much. You started with the same plan and boarded the bus at the correct time. While your bus is stopped at one of the stops you decide to get off to determine if a there is a train from point A to B since you want to go somewhere in a month from now by the train from point A to B. You could check this online through your phone or after reaching home. After getting the details you found that another bus that stops at the point where you got off comes at 5:30 PM and it will reach the movie at 6:30 PM. This is what cohesion is.

Just catching a bus which goes to the movie theatre in time makes your life easier and you could accomplish your objective without any risk. Since, you have deviated from accomplishing your task by adding unnecessary goals it failed your actual goal. So, cohesion should be very strongly interrelated. Unrelated steps of code in a single module deviates the module in accomplishing the actual goal.

Returning from layman terms

According to software principles, a Functional Cohesion is a good option where the module contains elements that perform exactly one function or achieves a single goal and Coincidental Cohesion is a very bad manner where the module performs multiple unrelated tasks. This may occur if modules are not properly designed. Sequential Cohesion is also a good option where the output of one element in the module serves as input for another element. A module with sequential cohesion is essentially an Abstract Data Type (ADT) and all the benefits of ADT use are obtained when this type of cohesion is employed in a single module. Such a module operates on a single data structure.

Content Coupling is a very bad option where one module directly references the content of another module. For example Module x branches to a local label in Module Y or module x modifies a statement of module Y. Such modules are inextricably linked to each other and so this is a very dangerous coupling. Data Coupling is a very good option where two modules are data coupled if all arguments are homogeneous data items. If a data structure is passed then the called module must use all the fields. Data coupling is a desirable goal. Modules which are data coupled are easier to maintain since changes in one is less likely to cause regression fault in another.

Note: There are other types available. For more details, refer to any software engineering principles articles.

So the concept of Classes and Objects help us design a system where all the entities are loosely coupled and are use tight cohesion principles. An object can be thought of as a module exhibiting Sequential Cohesion but also having other properties as well.

Objects support data hiding and encapsulation but have an extended characteristic called inheritance. The basic idea of inheritance is that new data types can be created from existing types without creating the types from the scratch.

With the OOP principles you can design a system which is loosely coupled and with strong cohesion by classes and inheritance. Each class is supposed to encapsulate related data and methods and provides public methods or constructors to provide the values and returns the output hiding its code for achieving the functionality.

SOLID principles help us make the system easy to enhance without changing the existing code which is going to break the other class functionalities. For example, my mobile phone has a charger which can be connected to my laptop or my Car USB port and by attaching a plug to the charger USB, I can use the same charger to connect to the electric power. This provides me the ability to attach or detach a particular part of the charger and still it works without affecting the functionality. I do not need to buy another charger for connecting to power and one for connecting to my laptop. Ideally, our software should behave like my mobile charger.

Now, let us see what each principle of SOLID says with a simple example. Usually we end up refactoring the code rather than building the code from scratch. So, here I will  use an existing project and see how to refactor it according to SOLID principles.

Single Responsibility Principle

At any instance of time your class should support only one purpose and if you change your class anytime that should be only for that one purpose. If your class needs to be changed for any other reason, then you might have to distribute the responsibilities to multiple classes i.e. divide the class into multiple classes.

Let us open the existing project. This project has a console application as a client and there are initially two class libraries. One is for business logic classes and the second one is a business entity class.

My entities class looks like below

using System;
using System.Collections.Generic;
using System.Linq;
using System.Text; 
namespace BusinessEntities
{
    public class Employee
    {
        public string EmployeeId { get; set; }
        public string EmployeeName { get; set; }
    }
}

My business logic class looks like below


using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using BusinessEntities;

namespace BusinessLayer
{
    /// <summary>
    /// Class to get Employee Data, Format and display the data.
    /// </summary>
    public class EmployeeData
    {
        /// <summary>
        /// Method to get the Employee data from Database
        /// Return the data to the calling method.
        /// </summary>
        /// <returns></returns>
        private IList<Employee> GetEmployeeData()
        {
           
var employees = new Employee[1];
            employees[0] = new Employee { EmployeeId = "111", EmployeeName = "John Sherman" };
            return new List<Employee>(employees);       
        } 
        private string FormatEmployeeData()
        {
            Console.WriteLine("Getting Employee Data...");
            var employessList = GetEmployeeData();
            Console.WriteLine("Formatting the Employee details...");
            return string.Format("Employee Name of Employee ID {0} is {1}",
                                                   employessList[0].EmployeeId, employessList[0].EmployeeName);
        }
        public void DisplayEmployeeData()
        {
            GetEmployeeData();
            string empData = FormatEmployeeData();
            Console.WriteLine("Displaying Employee Data as below:");
            Console.WriteLine(empData);
        }   
    }
}

My console application program file looks like below

using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using BusinessLayer; 
namespace SolidPrinciples
{
    class Program
    {
        static void Main(string[] args)
        {
            var employeeData = new EmployeeData();
            employeeData.DisplayEmployeeData();
            Console.ReadLine();
        }
    }
}


Let us execute the program first to see what the program is doing.

1.gif

My program is basically getting the employee data from the database and then formatting the employee data and then presenting the data in the UI.

So, let us observe the business class EmployeeData more closely. You would observe that the class is performing the following tasks:

  1. Getting employee data -This method gets data from the database for all the employees.
  2. Formatting employee data - This method formats all the data. You can consider this method as something that processes business logic.
  3. Displaying employee data - This method displays the formatted data in the UI. You can consider this method as something that binds the data to the UI controls.

So, the class has 3 different responsibilities whereas it is supposed to be handling one responsibility.

2.gif

So, now what we are going to do is to remove the first task, getting employee data, which is supposed to be in a data access class. So, we create a new class and move the GetEmployeeData method to the data access class as follows:

public
class DataAcess
{       
   public IList<Employee> GetEmployeeData()
   {
       var employees = new Employee[1];
       employees[0] = new Employee { EmployeeId = "111", EmployeeName = "John Sherman" };
       return new List<Employee>(employees);
   }
}

And also we want to format the data in a different class whose responsibility is only to format the data obtained from DataAccess. So, I am going to create a new class DatsFormatter and move the method to format into this class.

public
class DataFormatter
{
    public string FormatEmployeeData()
    {
        Console.WriteLine("Getting Employee Data...");
        var employessList = new DataAcess().GetEmployeeData();
        Console.WriteLine("Formatting the Employee details...");
        return string.Format("Employee Name of Employee ID {0} is {1}",
                                                   employessList[0].EmployeeId, employessList[0].EmployeeName);
    }
}

Now we also want to move the method to display employee data to a new class so that this class only has the single responsibility of showing formatted employee data; see:

public class DisplayEmployee
{
    public void DisplayEmployeeData()
    {
       string empData = new DataFormatter().FormatEmployeeData();
       Console.WriteLine("Displaying Employee Data as below:");
       Console.WriteLine(empData);
    }
}

Now our original business class has no methods and we create a method DisplayEmployeeData() and call the EmployeeDataMethod of the class DisplayEmployee. Now this class only has the one responsibility to call the method of the class DisplayEmployee.

public class EmployeeData

   public void DisplayEmployeeData()
   {
      new DisplayEmployee().DisplayEmployeeData();
   }
}

Now what I am going to do is to move each class into a separate file to make the code look neat and modification of the code becomes simpler; see:

3.png

Now each class has only one responsibility. EmployeeData creates an instance of DisplayEmployee to display the employee details. DisplayEmployee to get the data from DataAcess by creating an instance of DataAccess and getting the employee details and formats the employee details.

Now we have 4 different files available in the Business Layer. Now to simulate the project environment that we work at the office, I am going to create another class library project for DataAccess and move the DataAccess class to the new class library.

Now the final structure of the project is as below:

4.gif

Now the business Entities remaining the same, the BusinessLogicLayer is below:

5.gif

And the class code is as below

using System;
using DataAccessLayer; 
namespace BusinessLayer
{
    public class DataFormatter
    {
        public string FormatEmployeeData()
        {
            var dataAccess=new DataAccess();
            Console.WriteLine("Getting Employee Data...");
            var employessList = dataAccess.GetEmployeeData();
            Console.WriteLine("Formatting the Employee details...");
            return string.Format("Employee Name of Employee ID {0} is {1}",
                                 employessList[0].EmployeeId, employessList[0].EmployeeName);
        }
    }
}
 
using System; 
namespace BusinessLayer
{
    public class DisplayEmployee
    {
        public void DisplayEmployeeData()
        {
            string empData = new DataFormatter().FormatEmployeeData();
            Console.WriteLine("Displaying Employee Data as below:");
            Console.WriteLine(empData);
        }
    }
}

using System.Linq;
using System.Text;
 
namespace BusinessLayer
{
    public class EmployeeData
    { 
      public void DisplayEmployeeData()
      {
          new DisplayEmployee().DisplayEmployeeData();
      }           
    }
}

Client code remains the same

using
System;
using System.Collections.Generic;
using System.Linq;
using
System.Text;
using BusinessLayer; 
namespace SolidPrinciples
{
    class Program
    {
        static void Main(string[] args)
        {
            var employeeData = new EmployeeData();
            employeeData.DisplayEmployeeData();
            Console.ReadLine();
        }
    }
}

DataAccesslayer looks as below

6.gif

Code for DataAccess class is as below

using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using BusinessEntities; 
namespace DataAccessLayer
{
    public class DataAccess
    {
        public IList<Employee> GetEmployeeData()
        {
            var employees = new Employee[1];
            employees[0] = new Employee { EmployeeId = "111", EmployeeName = "John Sherman" };
            return new List<Employee>(employees);
        }
    }
}


Let us run the code again to see the output and it should be the same, as in:

7.gif

This is the end of the Single Responsibility Principle and you can find the code for this project in the folder named "SRP".

In the next article we discuss the OCP - Open/Closed Principle.

Hope you have enjoyed the first article on SOLID principles and your feedback on this is highly appreciated.

Next Recommended Readings