Architecting .NET apps

Question 1

What is a common problem in technology?

Answer 1

Learn something new and try to use it everywhere as at silver bullet.

Question 2

What is an expert as per the bridge saying?

Answer 2

Anyone can design a bridge that stands, but it takes an engineer to design a bridge that barely stands. Meaning you need to spend just enough resources to solve the problem.

Question 3

What are the reasons to select an architecture?

Answer 3

To: save money, effort and time. Instead of experimenting, challenging or improving the resume.

Question 4

What are the 5 common “thinkings” of the astronaut assessment (head in the cloud)?

Answer 4

• We must use this shining technology for this!
• Reinvent the wheel.
• Never ORM.
• Always code to an interface.
• Use all the patters from the GoF book.

Question 5

When is it allowed to introduce complexity to the application?

Answer 5

When it is JUSTIFIED. Ultimately we’re paid for solutions, not code.

Question 6

What is the Parkinson’s Law?

Answer 6

Work expands to fill available time.

Question 7

What the Parkinson’s law means in the software development when a developer has tight timeline?

Answer 7

He comes with smart and simple solution to meet the deadline. Developers somehow pull it off.

Question 8

How usually developers proves the Parkinson’s law? Do they made things simpler?

Answer 8

By leveraging the SIMPLICITY: the art of maximizing the amount of work NOT done.

Question 9

Why agile says about maximizing the work not done? Should we do everything right away?

Answer 9

You can wait as long as possible until the work is absolutely necessary (by then you might get more information that will help solving the problem).

Question 10

What are the 6 aspects of the “Do the simplest thing that could possibly work”?
Note: This is listed in the extreme programming wiki.

Answer 10

Get done sooner.
Easier to communicate;
Duplication is obvious, so the need for refactoring is clearer;
Tests are easier to write;
Code is easier to performance tune;
You feel less stress, which enhances all of above.

Question 11

What is the 80/20 rule?

Answer 11

Get 80% of the benefit by doing 20% of the work.

Question 12

What does the agile manifesto say?

Answer 12

Maximize the work not done;

Welcome changing requirements, even late in development.

Question 13

What are the 7 key lean principles?

Answer 13

1 - Eliminate waste;
2 - Amplify Learning;
3 - Decide as late as possible;
4 - Deliver as fast as possible;
5 - Empower the team;
6 - Build integrity in;
7 - See the whole.

Question 14

What are the specifics of lean principles?

Answer 14

1 - Not adding value to the customer is likely waste;
3 - When the choice is unclear, choose the simplest one;
4 - Ideal architecture varies by team;
6 - Automated testing.

Question 15

What is the main outcome of the minimum viable product?

Answer 15

The validated learning (user’s feedback). Is about answering questions.

Question 16

What is not part of a minimum viable product?

Answer 16

Usually expensive items, such as Scalability, beauty, code cleanliness, performance, security and if it even actually works.

Question 17

What is the project triangle? Which one is usually sacrificed for the sake of others?

Answer 17

A triangle containing quality, date and cost… increasing any of the edges, will increase the others. Quality is usually sacrificed in form of technical debits.

Question 18

What is the difference between hard and soft deadlines?

Answer 18

Hard: you have to hit the milestone due to, e.g make a published advertising
Soft: someone said so based on guess. The software architect usually explains to the non-tech people what are the implications.

Question 19

What are common layers in software architecture? What are their equivalent terms?

Answer 19

• presentation aka UI;
• service aka API;
• Domain aka Business Logic;
• Persistence aka Data Access (DAL)

Question 20

What is the difference between layer and tier?

Answer 20

layer is logical and tier is physical or virtual.

Question 21

What are the pros and cons of tiers?

Answer 21

• Pros: Scalability, Security, Uptime, Reusability.

- Cons: Performance cost and Increased complexity.

Question 22

How the layers are represented in .NET? What are the benefits?

Answer 22

Each layer is a csproj… business layer (and business layer.tests), data layer and presentation layer. The benefits are separations of concerns, abstract complexity, support testing via interfaces, minimize dependencies and enable reuse.

Question 23

How does the layers are represented in a very simple app (e.g console app)?

Answer 23

Maybe methods are a suffice separation for such application.

Question 24

What is the business logic layer?

Answer 24

Manages the behaviors that are triggered by the presentation layer. This layer shields the presentation layer from the persistence logic

Question 25

What are the business layer patterns? From less to most complex?

Answer 25

Transaction script (procedural), table module, active record (both data driven) and domain model (business driven).

Question 26

How the Transaction Script pattern work? When is it a good? What are the risks?

Answer 26

One public function per UI operation. Very simple but kind of procedural.
It is good for tight deadlines.
There is the risk of duplication and difficult to maintain overtime.

Question 27

What is the Table Module pattern? Is it legacy?

Answer 27

Classes represent tables (datasets and datatables). Used to be very popular. Yes, it is legacy approach - replaced by entity framework.

Question 28

What is the Active Record pattern? What are the pros and cons?

Answer 28

Each class instance represents a row (similar to dataset?). Class knows how to persist itself and contains business logic.

pros: simple, good for simple domain and CRUD applications.
cons: rigid domain (tied to DB), leads to swiss switchblade classes, low cohesion, hard to test and tricky transactions.

Question 29

What is the domain drive design? What are the pros and cons.

Answer 29

It is about understanding the business and terms used not bothering with how the persistence layer will look like.
Pros: Manage complexity, leverage design patterns, speak business logic, abstract ugly db schema, compliments large teams and it is reusable.
Con: Learning curve, time-consuming design, long-term commitment (investment upfront), DB mapping overhead.

Note: Suggested book: Eric Evans - Domain Driven Design

Question 30

What is object relational mismatch?

Answer 30

It is when you classes do not match the database tables.

Question 31

What are the .NET tooling for each business layer pattern?

Answer 31

1 - Transaction script: None;
2 - Table Module: Data tables and datasets;
3 - Active Record:EF / Linq to SQL / Castle Active Record;
4 - Domain Model: None. Note: DB is separate concern;

Question 32

What is the author’s heads-up about Interactions?

Answer 32

• He says that only logically interrelated classes should interact in the business.
• The Service Layer should handle all other interactions.

Meaning that complex interactions should happen in the service layer.

Question 33

What are the main responsibilities of Service Logic Layer?

Answer 33

• Acts as an intermediary between the presentation layer and the business logic layer.
• Coordinate the domain, meaning complex interactions will be abstracted.
• Can be seen as a facade (simplify complex calls and make the classes loosely coupled).
• Takes requests from one layer and sends to another (taking care of calling order).
• Can centralize handling cross-cutting concerns such as logging, error-handling and security (authe and autho).
• Can handle searching, notifications to listeners, bindings and managing transactions (e.g only send the email if all the other business logic succeeds).

Question 34

Why do we say that the service layer should be thin?

Answer 34

Because it should only make it easier to the presentation layer to call complex routines in the business and it must delegate most of the work to the business objects. Also data validation.

Question 35

What is a real-world example of service layer when we look to the restaurant?

Answer 35

The waitress interfaces (she and the menu are the API) customers orders to the kitchen. So that customers do not know about complexities, they simply get their orders in the end.

Question 36

What is the real-world example of the service layer when we look to a boss?

Answer 36

He accomplishes tasks through others.

Question 37

When to use a Service Layer? Does it have to do with reusability and complexity?

Answer 37

When you have multiple UIs, so that you can centralize some calls in the Service Layer. Multiple consumers of business logic. And when you have complex interactions among domain objects.

Question 38

What are the pros and cons of deploying the Service Layer as a service or a shared library? Which one is the default?

Answer 38

Webservice:
1 - Clients are independent of the service.
2 - Can scale.
3 - Centralized deployment.

Shared libs (Default to this):
1 - Native code in the process.
2 - No serialization overhead.
3 - works offline. always available. no public exploit. Simplest thing that works.

Question 39

What should the service layer return?

Answer 39

1 - Data Transfer Objects (DTO - aka class with no methods);
2 - Copy of domain entities without behavior;
3 - Simplest thing is to expose the Domain entity.

Question 40

What is a DTO? When to use it?

Answer 40

Object that carries that across an application’s boundaries with primary goal of minimizing round trips. Avoids coupling between UI and domain layers.
To solve circular references. Efficiency by creating DTOs with only the necessary data and easy structure.

Use them where it makes sense. Useful but optional.