Session04 Additional Info

1. Should I have an IEntityService and then EntityService for each of my entities?

Not necessarily for every entity — only if it makes sense.

  • The Application Layer should expose use cases — not just CRUD logic for each entity.
  • If an entity has business logic or interactions that need orchestration (e.g., validations, aggregations, calling repositories, etc.), then yes, create a service.
  • Otherwise, for basic operations, directly using a repository (via a unit of work or interface) from the use case handler might be fine.

Absolutely, yes. In fact, that’s expected in a Clean Architecture setup.

Examples:

  • A ReportGenerationService that combines bookings, customers, and payments.
  • A TokenService for authentication tokens.
  • A CurrencyConversionService that hits an external API.
  • A NotificationService that sends emails or SMS.

👉 As long as these services live in the Application Layer and follow dependency inversion (i.e., they depend only on interfaces, not implementations), you’re doing great.

3. Is it necessary to have an interface for each service?

🔹 What’s the Difference Between Services and Repositories?

Aspect Service Repository
Layer Application Layer Domain Layer (interface), Infrastructure Layer (implementation)
Responsibility Orchestrates business logic / use cases Data access abstraction
Focus Coordinates multiple domain/repo operations, validation, business flow Fetching/storing data for a specific entity
Example PlaceOrderService, ReportService ICustomerRepository, IOrderRepository

What does init mean?

init is an access modifier for properties that allows you to set a property only during object initialization, but not after.

Should I use class or record for DTOs in Clean Architecture?

🔵 Short answer:

Use record for DTOs when possible — it's clean, immutable by default, and semantically perfect for data transfer.


🔍 Why record is a great fit for DTOs

Feature record class
Immutable by default ✅ (with init) ❌ (need manual setup)
Value-based equality ❌ (ref-based by default)
Concise syntax ❌ (more boilerplate)
Use for data containers ✅ (perfect fit) ✅ (but more verbose)
Custom behavior/logic ❌ (less suitable) ✅ (better for rich behavior)

But when should you prefer class?

Use class if your DTO or model:

  • Needs to be mutable after creation
  • Has to interact with legacy APIs/libraries
  • Needs inheritance or polymorphism (not well supported in record)
  • Has rich behavior (logic, methods, validation, etc.)

For example, in the Domain Layer (Entities, ValueObjects), you'll usually stick to class — because that's where behavior lives.

What Conditions Make an API RESTful?

Key Principles of REST:

  1. Statelessness:

    • Each API call must contain all the information the server needs to fulfill the request (no session state). Each request is independent.
  2. Resource Identification:

    • Resources (e.g., customers, orders) should be identified using URIs. Use nouns in URIs, not verbs.
  3. HTTP Methods:

    • Use standard HTTP methods to represent actions:
      • GET: Retrieve a resource.
      • POST: Create a new resource.
      • PUT: Update a resource entirely.
      • PATCH: Update a resource partially.
      • DELETE: Remove a resource.
  4. Use of Standard Status Codes:

    • Return appropriate HTTP status codes (e.g., 200 OK, 201 Created, 404 Not Found, 500 Internal Server Error).
  5. HATEOAS:

    • (Hypermedia as the Engine of Application State) - Provide links to related resources within the responses.

Multiple GET Methods in One Controller:

  • Yes, you can have multiple GET methods in one controller. The key is to differentiate them based on routes and parameters.
  • For example:
[ApiController] 
[Route("api/[controller]")] 
public class CustomerController : ControllerBase {     
	
	[HttpGet("{id}")]     
	public IActionResult GetCustomerById(int id) { /*...*/ }      
	
	[HttpGet]     
	public IActionResult GetAllCustomers() { /*...*/ }      
	
	[HttpGet("{id}/orders")]     
	public IActionResult GetCustomerOrders(int id) { /*...*/ } 
}
  • Routing: ASP.NET Core uses route templates to differentiate these actions. The combination of route parameters, query strings, and action names can help separate the GET requests.

List-like stuff in C#

Absolutely, let’s go over the main “list-like” data types in C#. They all serve similar purposes—holding multiple items—but differ in functionality, performance, and use cases. Here’s a detailed breakdown:

🔷 1. IEnumerable

  • Namespace: System.Collections.Generic
  • Most basic "list-like" abstraction.
  • Read-only (forward-only iteration).
  • You can use foreach on it.
  • Doesn’t support indexing (no .Count, no [i]).
  • Often used as the return type to expose a stream of data without giving full collection control.

Example:

IEnumerable<int> numbers = GetNumbers(); // Lazy-loaded maybe
foreach (var num in numbers)
    Console.WriteLine(num);

💡 Ideal when:

  • You want to return a sequence without exposing modification.
  • You’re using LINQ chains.
  • You’re returning data from a database query.

🔷 2. ICollection

  • Extends IEnumerable.
  • Adds Count and Add/Remove/Clear methods.
  • Still abstract—List and HashSet implement it.

💡 Useful when:

  • You want to expose a collection that can be modified (e.g. Add or Remove).
  • You care about the Count.

🔷 3. IList

  • Extends ICollection and IEnumerable.
  • Adds index access: list[0] etc.
  • Think of it like a mutable array with dynamic size.

💡 Use when:

  • You want ordered collection with indexing.
  • You need to insert, remove, or replace items at specific positions.

🔷 4. List

  • A concrete class (not interface).
  • Implements IList, ICollection, IEnumerable.
  • Backed by an array (auto-resizes).
  • Fast read and write.
  • Supports Add, Remove, Insert, IndexOf, etc.

Example:

var list = new List<string>();
list.Add("One");
list.Add("Two");
var second = list[1]; // "Two"

💡 Go-to general purpose collection.


🔷 5. IReadOnlyCollection & IReadOnlyList

  • IReadOnlyCollection: Just Count and IEnumerable.
  • IReadOnlyList: Adds indexing without modification.
  • Used to expose lists safely (read-only).

💡 Used when:

  • You want to return a list, but prevent any changes.

🔷 6. Array (T[])

  • Fixed-size.
  • Fastest for indexing.
  • Cannot change size.
  • Implements IList (via Array).

Example:

int[] numbers = new int[5];
numbers[0] = 42;

💡 Great for performance-sensitive or fixed-size data.


🔷 7. ObservableCollection

  • For WPF/Blazor/WinForms data-binding.
  • Notifies UI when items are added/removed.
  • Implements IList.

💡 Use in UI apps when the view needs to react to collection changes.


🔷 8. HashSet

  • Unordered, no duplicates.
  • Implements ICollection, not IList.
  • No index access.

💡 Best for fast membership checking (contains x).


🔷 Summary Table

Type Indexing Modifiable Ordered Duplicates Use Case
IEnumerable ✔️ ✔️ Basic read-only stream or LINQ chaining
ICollection ✔️ ✔️ ✔️ Expose modifiable group of items
IList ✔️ ✔️ ✔️ ✔️ Full-featured list abstraction
List ✔️ ✔️ ✔️ ✔️ Standard go-to collection
IReadOnlyList ✔️ ✔️ ✔️ Return data safely from service or repo
Array (T[]) ✔️ ✔️ ✔️ ✔️ Fixed-size, performant indexing
ObservableCollection ✔️ ✔️ ✔️ ✔️ UI binding for collections
HashSet ✔️ Fast lookup, uniqueness enforcement

If you tell me the scenario (e.g. search results, modifying a cart, populating a drop-down), I can recommend the best type for it. Want that?