Redundant id key in json schema?

Why is there this additional id key under citationItems? Isn’t that redundant?

Or redundant with a purpose; easier processing?

Example:

{
   "schema":"https://resource.citationstyles.org/schema/latest/input/json/csl-citation.json",
   "citationID":"item1",
   "citationItems":[
      {
         "id":"item1",
         "itemData":{
            "id":"item1",
            "type":"book",
            "title":"test title"
         }
      }
   ]
}

_Originally posted by @bdarcus in https://github.com/citation-style-language/schema/pull/233_

[bwiernik]

@bdarcus I think the best way to approach this would be to leave citekey just in the data—Essentially as a human-readable alias for id.

[fbennett]

(I think this is relevant, but correct me if I’m wrong.) This data shape is a processing outcome. The citationItems objects initially carry just the id, locator, and label, and a processor will fetch itemData to the object as a pointer based on the id. In some contexts (bibliographies, sort-key generation) item data is handled as an array of objects, and the id needs to be in there. To remove id from the itemData segment in citations would require cloning the object, which would break pointers, raise memory consumption a bit, and run the risk that multiple metadata representations of the same item would creep into documents. Removing id from the parent citationItems objects would be to delete the id received on input, which might not be a good idea.