The goal of enumerating each well-known graph property and family conflicts with the goal of being a light-weight dependency. Even though this package does not have almost any executable code, defining a great number of new types could still be taxing on the compiler, I guess.
It might make sense to:
-
Split up the abstract supertype from the definitions of particular properties. Then loading the abstract supertype would not imply loading all the properties.
-
Split up some of the particular property definitions from some of the others. Then loading a single concrete property would not imply loading all properties.
The goal of enumerating each well-known graph property and family conflicts with the goal of being a light-weight dependency. Even though this package does not have almost any executable code, defining a great number of new types could still be taxing on the compiler, I guess.
It might make sense to:
Split up the abstract supertype from the definitions of particular properties. Then loading the abstract supertype would not imply loading all the properties.
Split up some of the particular property definitions from some of the others. Then loading a single concrete property would not imply loading all properties.