Your support for our advertisers helps cover the cost of hosting, research, and maintenance of this FAQ
It is very suitable for representing rectangular data (row-and-column data, see the topic ‘Data-style applications’ in question D.14 on ‘What is my information? DATA or DOCUMENT?’) including that destined for (or coming from) spreadsheets, label/value sources, relational databases, and other sources, particularly where the structure is not deeply nested.
It is not suited to the storage of information using Mixed Content (see the topic ‘Document-style applications’ in question D.14 on ‘What is my information? DATA or DOCUMENT?’) — the normal ‘document’ XML format where text and markup are arbitrarily intermingled to any depth, such as in traditional XML applications, eg DocBook, TEI, JATS, question A.5 on ‘What is HTML?’, etc. In such cases, JSON becomes unwieldy and ceases to be human-readable.
There are ongoing discussions at
suggestions for standardising definitions, paths, and transformations;
Google’s attempt to use JSON for Mixed Content;
IBM’s JSONx attempt to represent JSON in XML;
See @SwiftOnSecurity: ‘JSON is completely unreadable for configurations. It's a programmatic data transfer format. Don't do this to your users. Use XML.’