On Monday, June 25, 2018 11:35:35 AM EDT Michael Kay wrote:
That fallacy is about to be exposed. JSON is "simpler" to use because up
till now it has been used for simpler things.
JSON *is* simpler to use for simple things, no doubt about it. Mainly
because it is a much closer fit to the data models of conventional
programming languages.
In JSON, each value is either a list of values or a set of key/value pairs.
When the data model fits JSON, it's easy to use. It becomes hairy when values
are html, xml, or json, because of escaping of double quotes. Most programming
languages have matching data structures, so it's easy to use JSON to exchange
them via an API. Also, JSON is fairly human readable, unless there is a lot of
escaping.
On the other hand, when your data are documents or when you don't mind writing
a more complex (de)serialization code for your data structures to XML, XML
provides better (de)serialization fidelity, more human readable than JSON,
especially when values have few escaped angular brackets. Once you get over
the hump of getting your data into XML, all the tooling that you can use for
validation and transformation. That's the advantage of XML.
