# Part1

This document covers long-term vision. A shorter-term, minimal-viable-product is discussed here.

This covers one set of implementation choices. Alternatives are discussed here.

Business value, rationale, justification are discussed here.

Related works are discussed here.

Internal design is discussed here.

To demonstrate the framework's capabilities, we will walk through building a realistic subset of Stack Overflow functionality.

First, run the maven archetype.

mvn archetype:generate \
  -DarchetypeGroupId=com.klass \
  -DarchetypeArtifactId=klass-maven-archetype \
  -DinteractiveMode=false \
  -DgroupId=com.stackoverflow \
  -DartifactId=stackoverflow \
  -Dversion=1.0.0 \
  -Dpackage=com.stackoverflow \
  -Dname=StackOverflow

This command creates a new project in ./stackoverflow/. The project compiles and runs, but doesn't yet do anything interesting.

The model was generated in ./stackoverflow/stackoverflow-domain-model/src/main/resources/com/stackoverflow/stackoverflow.klass. It is nearly empty except for the package declaration.

An Klass model consists of classes, enumerations, associations, projections, and services.

The classes, enumerations, and associations make up the data model. From the data model, Klass generates a database schema, POJOs, DTOs, example data, etc. We will define the Question and Answer classes, as well as the one-to-many association between questions and answers.

Services are traditional http or rpc services. We will define a service to get a Question by its id.

Projections map services to data. They answer how much data a service reads or writes; conceptually how far the service walks through the object graph. For example, when we fetch a Question by its id, should we return the title and the body, or just the title? Should we include the question's answers in the response?