A schema is a blueprint for an attestation. It describes the various fields an attestation contains and what their data types are. Anyone can create a schema in the registry, and once created, schemas can be re-used by anyone else.
Schemas are stored in the registry as a string value that describes the various fields. For example, to create attestations that describe a person, we can create a schema as follows:
As you can see from this example, a schema is a comma-separated list of tuples, composed of property name and datatype.
Schemas with Nested Data
Many schemas will involve some kind of nested data. To create a schema with nested data, place the nested data in round braces, followed by the property name, for example:
As well as the schema string, schemas have some metadata stored with them:
Schema Contexts / Shared Vocabularies
Schemas have a field called "context", usually a link to some shared vocabulary. This is an important field that gives semantic meaning to the schema, allowing consumers of the attestation to understand exactly what the fields in the schema represent. This removes ambiguity and allows reputation protocols to build powerful semantic graphs when indexing attestations.
For more information on this property and how important it is, see the Linked Data page.
Counterfactual Schemas
Schema IDs are unique to the schema content, and are created from the keccak hash of the schema string. This means that schemas can be created counter-factually, allowing for things like circular relationships between schemas.
Experimental Features for Advanced Schemas
These features are advanced and experimental schema features that allow you to create schemas with sophisticated properties. Feel free to get in contact if you have any questions.
NOTE: caution should be taken with these features as they are experimental and likely to change, and may not be supported by all indexers!
How to create related schemas
Sometimes, you may wish the consumers of your attestation to know how the attestations relate to other attestations. To do this, you create a relationship. So, for example, to create an attestation of a player that is a member of a team, one would first create a Team schema, and then you would create a Player schema with a canonical relationship field, denoted by curly braces:
string pseudonym, string dateJoined, { isMemberOf Team 0xa1b2c3 }
… where isMemberOf is the relationship type, Team is the schema name, and 0xa1b2c3 is the schema ID. This indicates that any attestation based on the above schema is expected to be linked to some other Team attestation via a relationship attestation that links them together.
This approach reduces redundant attestations, and allows for more fine-grained and reusable schemas, contributing to a growing standard library of well-known and widely used schemas.
Similarly, to create a one-to-many canonical relationship field, you would use the syntax:
string firstName, string lastName, [{ isResidentAt Place 0xa1b2c3 }]
As anyone can create links between different attestations, including a canonical relationship in your schema can help the consumers of your attestations understand what relationships you intended on being there. As opposed to relationships that may be created by third parties.
See the page on Linking Attestations to understand how to link attestations after they are created.
A note on relationship attestations
The examples above use what is called a "relationship attestation", meaning any attestation based on the special Relationship schema. The relationship schema conforms to the following structure:
This Relationship schema exists as a first-class citizen of the registry, and attestations that are based on this schema are used for linking other attestations together. The subject field is the attestation that is being linked to another attestation, the predicate field is a name that describes the typing of the relationship, and the subject is the attestation being linked to.
Examples of relationship attestations are:
0x46582... "isFollowerOf" 0x10345...
0x31235... "hasVotedFor" 0x52991...
0x74851... "isAlumniOf" 0x31122...
Anyone can create any kind of relationship between any attestation and several other attestations, allowing for the emergence of an organic folksonomy. However, it also makes canonical relationships important to define in the schema. Otherwise, there will be ambiguity between which relationship attestations were intended by the attestation issuer, and what were relationship attestations that were arbitrarily added later by third parties.
Schemas can also inherit from other schemas, another way that Verax reduces redundant schema data and promotes re-usability. To inherit from another schema, add the parent schema ID at the start of the schema string preceded by the @extends keyword, e.g.:
This will tell indexers to look up the schema referenced by the extends keyword, and concatenate its schema string with the schema string in this schema. Note that any conflicting field names will be overridden by the last previous definition. So, for example, if a field name exists in a parent schema and a child schema, the field definition from the child schema will be used. Also, schemas can only inherit from one parent at a time.