First of all, thanks for this package! It’s a good step towards versioning and interoperability of messages.
While this may be somewhat subjective, I find the name identifier a bit confusing. Identifier suggests a unique id for an individual message. Think of the response you get from AWS when you publish a message:
I think what you meant is not the identifier but the type of the message: as in a typing system, each (product) type is the combination of properties; in this case the keys in the message payload.
What about renaming identifier to type? Of course, this would be a BC break.
Just because it is an "identifier", it does not mean it is a "unique identifier".
Im on the fence. You are correct that "type" os more clear. But is it worth the BC break?
First of all, thanks for this package! It’s a good step towards versioning and interoperability of messages.
While this may be somewhat subjective, I find the name
identifier
a bit confusing. Identifier suggests a unique id for an individual message. Think of the response you get from AWS when you publish a message:I think what you meant is not the identifier but the type of the message: as in a typing system, each (product) type is the combination of properties; in this case the keys in the message payload.
What about renaming
identifier
totype
? Of course, this would be a BC break.