Hey all,

Creating this issue to discuss small but important changes to the GraphQL Getting Started Experience.

When you visit graphql.org landing page, you land on this hero banner.

The top banner here speaks to a client audience (mostly frontend focussed) where benefits of GraphQL for consumption is highlighted.

• Describe your data
• Ask for what you want vs Expose what data you want
• Get predictable results

This needs to address both the frontend and backend audience where the highlights need to balance out what is beneficial for the backend producer.

When you click on Get Started here, it takes you to graphql.org/code.

The hero banner here talks about Code using GraphQL

GraphQL tooling has evolved to allow a different pattern of building GraphQL servers by introspecting a data source (a database or an existing API like an OpenAPI spec) and auto-generating a GraphQL API without writing any boilerplate CRUD code, essentially acting as a compiler. Some examples of such projects are PostGraphile, Hasura, Tuql, Neo4j GraphQL, IBM’s OpenAPI to GraphQL, Mesh’s source handlers, etc.

In summary, two broad approaches to building a GraphQL API have emerged,

• DIY GraphQL: define a GraphQL schema and write resolvers
• Use tools that introspect the underlying source to generate the GraphQL API/schema.

We need to represent these two broad approaches to developers who are looking to get started with GraphQL on the backend.

Happy to start a PR on this, but would love to get more thoughts before I do that.