To lazy load references we need to keep references separate from the text of an article.
Although we could use REST APIs for this, this seems like it would be a useful feature of the Cite extension, apps and mobile web being obvious first consumers.
We thus need to store them somewhere, potentially the ParserOutput.
- A yes/no answer to the question in the title "Can we store a JSON blob of references data in ParserOutput"
- If no, a proposed alternative.