 the data that you need, then you have custom endpoints because the problem is because the impact that you see by default will not change the response that you get when you interact with it, and by the documentation return. So, next we're going to interact with all, interact with all. So interact with all is actually one of our things. First of all, it's a great language for the client to describe the shape of the data that they need to go from the server. Second thing is the type system that the client and the server have a shape of the memory of what all the experience are in this country. The client is the one client to the server to translate the queries from the client into a distance from the client. So it's divided by 3 to 13, the fourth time the problem is released. It's the last time that the client doesn't make any assumptions about the database or the data store that you're using. It is 3 bones, 2 bones, even the rest of the data. It's a language that comes in handy. Finally, it's got a real code in most of the language to do it. So we can add it into the GraphQL board. So many of Graph in GraphQL has actually nothing to do with the Graph. It's actually one of the types that we can technically model our distance between us and the Graph. So this is a diagram. So this is like a library system. We have books and options. So if you look at the top book, the one with the dash, the one with the book has the title. So it has the edge, the column, the column, the script, and the column. And it has authors. So the edges there are 100 relationships, between the authors and the authors. And GraphQL has to extract trees from this graph of relationship. So if I follow one hospital tree to my tree, starting from the top book, you fill in the graph to follow the outgoing arrows. So that's one tree. So only if you use that graph, it is going to graph. So this is the same idea. Graph. So this is the actual GraphQL tree. Start on the left to the GraphQL tree. And on the right is the response to the GraphQL service. So if you look at this, instead of our concept on the shape of the GraphQL tree, we have the response to the GraphQL tree. And this is the GraphQL tree. So the key idea here, is that you need to help the client to serve the server. The client can choose to specify only the degree that it needs. And the server will do what's necessary to optimize on the server side. Related to the area. So the client is more of a closer to the server. And GraphQL server is just like the best server for an RTC server. It accesses GraphQL trees from clients and returns to the company. And in the end, it sees GraphQL service and the company to handle the customers. And it's completely affected the way. So again, it applies the same case as RTC and RTC. The meaning or the underlying. So this is an example. This is the URL. You have to follow up. The GraphQL ID is called GraphQL. It says here it's like a postman or what ID for GraphQL is listed. GraphQL is listed. So I'm not here to be anti-operating. I'm just going to put a phrase. So I'm going to say you want to return posts. So I'm going to put a phrase. So here we're treating posts the IDs for all our posts. I'm just going to want to treat for a different attribute. I can see what this means. It returns a post. It returns an array of types of posts. So let's look at one. Post and return. The post is used for these fields. Let's characterize the author. The author is a type user. So you treat the extract of the complex graph of relationships. What is the domain? The single. And one other thing is one thing about GraphQL is just that there's a line escape from both the server and the client. And the client will perform client-side operation. The app will maintain the environment that the client can go even before he's standing at the server that will be a very important problem. So wow, the first time I saw this I was like, oh, shit. This is the thing. And when you compare the user experience of GraphQL to the part you're making it's clearly a GraphQL and the rest of the user experience. And when you think about it, having a great detail on the user is actually a good option. Especially in the product. And how is it possible? I don't know why GraphQL needs to be possible. So if you use the artist between most of these types of fields in the world or if you use the type of those types of fields in the world I really don't know about other types of fields in the world. So the screen language the object specifically that will in the working type form will be a first object because it will be a title, author, and comment. And each field has a type that will be a stereotype and a stereotype to the team or to be an other customer type to use that in the company. So the explanation comes to that in the stand-up now. We also have to define the roots, possibilities of our life And it's a full type system where there are more advanced versions but none of this in the DCS type unions and some types and so on that is very important. So how does GraphQL solve the real risk to the product? But what is the scene around it? You don't have to need multiple response to get all the GraphQL solution to it. It's kind of specified that you really have to have this and it's not reflective. Okay, by adding between your business learning you get a lot of feedback from the GraphQL feedback high-side variation and from the GraphQL. So now, to make it more concrete let's look at the actual GraphQL server between you and me. So, to create a GraphQL server if you can it has to be my description of the different learning from the application. And this is all the questions. We start with GraphQL, we're going to move to the data problem. And let's be careful to be used in both and express.