There are several problems related to the design of the business logic of the annotation components. The most important is how much of the logic about the morphology of different lexical items should be included (or not included) into the annotation component and from where and how an annotation component should get the missing information:
@balmas: we are going to have to have the domain component tell us what the [annotation components'] options are -- they will change depending on the language, particular word, etc.
@balmas: I think it's an interesting question where the knowledge about which component to import should go. On the one hand, it's appealing to have that all encapsulated in the annotation package, but on the other hand, the domain knowledge about how a specific data object should be annotated probably belongs with the domain component. I think this will become clearer as we design the GraphQL API for the mutations.
There are several problems related to the design of the business logic of the annotation components. The most important is how much of the logic about the morphology of different lexical items should be included (or not included) into the annotation component and from where and how an annotation component should get the missing information:
@balmas: we are going to have to have the domain component tell us what the [annotation components'] options are -- they will change depending on the language, particular word, etc.
@balmas: I think it's an interesting question where the knowledge about which component to import should go. On the one hand, it's appealing to have that all encapsulated in the annotation package, but on the other hand, the domain knowledge about how a specific data object should be annotated probably belongs with the domain component. I think this will become clearer as we design the GraphQL API for the mutations.