Types

If you've used a statically defined language before - like Rust, Go, or Clojure then you'll already be familiar with strong type systems. You select the type - string, number, list, etc. - of data that a field will store. This improves memory allocation and access speed, as well as making sure that if you assign an incorrect type to the field an error is thrown. If you are unsure about typing a field, assign it an 'any' type - or ask us for help!

Valid Types

Type	Example
Strings	"Hello"
Booleans	True
Numbers	4.00
Structs (objects with typed fields)	{"foo": "bar"}
Arrays (ordered collections containing the same type)	[1,2,3]
Fixed-size Arrays	[1,2,3] (max 3 elements)
Any	JS objects, arrays, Python objects, etc

The any type designation can apply, appropriately enough, to any data type - it tells HASH to store the value as JSON and deserialize it at runtime. The generic any type can be used to simplify behavior key representations, but will result in slower simulation execution speeds than other type designations, so should be used sparingly.

All fields are nullable (which means they can be assigned a null value, or may exist with no value assigned) by default. You can turn off nullability, making a field non-nullable, which will improve simulation performance and memory utilization further. You can find the nullability setting by clicking the 3 dots.

For complex data types - lists, fixed sized lists, and structs, you must click the tree icon to assign the data types of members of the list or struct. You can find the setting for fixed-size list length by clicking the 3 dots.

Data type fields must be the same across behaviors. For instance if field foo in behavior A has type: number, field foo (assuming its the same field) in behavior B must have type: number.