Adapt Specification to Make the Schema System Optional

[zicklag]

Schema IDs should represent a hash unique to the schema, but that doesn't need to be a Leaf Schema formatted schema, it could be anything custom. The hash makes sure that we don't have a collision of schema IDs for different standards.

There should be a standard Leaf schema format, but it doesn't need to be prescriptive. If you don't understand the schema ID you can ignore it like any other component you don't know how to use.

This also makes it easier to adopt the Leaf protocol, because the schema format is one of the trickier pieces of it.

It also lets you use more common formats such as JSON for components, instead of Borsh, even though Borsh deserializing will be required for the core types such as entities.