Over the last 15 years, JSON has become ubiquitous on the web. Today it is the format of choice for almost every publicly available web service, and it is frequently used for private web services as well.
The popularity of JSON has also resulted in native JSON support by many databases. Relational databases like PostgreSQL and MySQL now ship with native support for storing and querying JSON data. NoSQL databases like MongoDB and Neo4j also support JSON, though MongoDB uses a slightly modified, binary version of JSON behind the scenes.
Below we’ll take a quick look at JSON and discuss where it came from, its advantages over XML, its drawbacks, when you should use it, and when you should consider alternatives. Let’s start with an example:
“worksWith”: [“Spantree Technology Group”, “InfoWorld”],
The structure above clearly defines some attributes of a person. It includes a first and last name, the number of times the person has logged in, whether this person is a writer, a list of companies the person works with, and a list of the person’s pets (only one, in this case). A structure like the one above may be passed from a server to a web browser or a mobile application, which will then perform some action such as displaying the data or saving it for later reference.
Why should I use JSON?
To understand the usefulness and importance of JSON, we’ll have to understand a bit about the history of interactivity on the web.
In the early 2000s, interactivity on the web began to transform. At the time, the browser served mainly as a dumb client to display information, and the server did all of the hard work to prepare the content for display. When a user clicked on a link or a button in the browser, a request would be sent to the server, the server would prepare the information needed as HTML, and the browser would render the HTML as a new page. This pattern was sluggish and inefficient, requiring the browser to re-render everything on the page even if only a section of the page had changed.
Fast-forward to the present day, and JSON is the de-facto standard for exchanging data between web and mobile clients and back-end services.
JSON versus XML
As noted above, the main alternative to JSON is XML. However, XML is becoming less and less common in new systems, and it’s easy to see why. Below is a version of the data you saw above, this time in XML:
<works_with>Spantree Technology Group</works_with>
Limitations of JSON
Although JSON is a relatively concise, flexible data format that is easy to work with in many programming languages, there are some drawbacks to the format. Here are the five main limitations:
- No schema. On the one hand, that means you have total flexibility to represent the data in any way you want. On the other, it means you could accidentally create misshapen data very easily.
- Only one number type: the IEEE-754 double-precision floating-point format. That’s quite a mouthful, but it simply means that you cannot take advantage of the diverse and nuanced number types available in many programming languages.
- No date type. This omission means developers must resort to using string representations of dates, leading to formatting discrepancies, or must represent dates in the form of milliseconds since the epoch (January 1, 1970).
- No comments. This makes it impossible to annotate fields inline, requiring additional documentation and increasing the likelihood of misunderstanding.
- Verbosity. While JSON is less verbose than XML, it isn’t the most concise data interchange format. For high-volume or special-purpose services, you’ll want to use more efficient data formats.
When should I use JSON?
If you’re writing software that communicates with a browser or native mobile application, you should use JSON as the data format. Using a format like XML is an out-of-date choice and a red flag to front-end and mobile talent you’d otherwise like to attract.
In the case of server-to-server communication, you might be better off using a serialization framework like Apache Avro or Apache Thrift. JSON isn’t a bad choice here, and still might be exactly what you need, but the answer isn’t as clear as for web and mobile communication.
If you’re using NoSQL databases, you’re pretty much stuck with whatever the database gives you. In relational databases that support JSON as a type, a good rule of thumb is to use it as little as possible. Relational databases have been tuned for structured data that fits a particular schema. While most now support more flexible data in the form of JSON, you can expect a performance hit when querying for properties within those JSON objects.
JSON is the ubiquitous, de facto format for sending data between web servers and browsers and mobile applications. Its simple design and flexibility make it easy to read and understand, and in most cases, easy to manipulate in the programming language of your choice. The lack of a strict schema enables flexibility of the format, but that flexibility sometimes makes it difficult to ensure that you’re reading and writing JSON properly.
You may have to do a little more work to work with JSON in strongly typed languages like Scala or Elm, but the widespread adoption of JSON means there are libraries and utilities to help you through all of the hardest parts. Choosing XML over JSON when building a new web service is a red flag likely to turn away talent, because it indicates a lack of interested in staying relevant.