Currently, database names are specified via string literals throughout the code base (wherever that database is used). As a result, determining the names of the database this application uses involves "searching (which also involves knowing what to search for)."
I propose that all database names be defined (I think of this as "registered") in a centralized place (e.g. a single Python module) and then import-ed from there where needed. That way, determining the names of the database this application uses will involve only checking that one file.
Currently, database names are specified via string literals throughout the code base (wherever that database is used). As a result, determining the names of the database this application uses involves "searching (which also involves knowing what to search for)."
I propose that all database names be defined (I think of this as "registered") in a centralized place (e.g. a single Python module) and then
import
-ed from there where needed. That way, determining the names of the database this application uses will involve only checking that one file.