keep your JSON in the mudroom

Sunday January 28th 2024 by socraticDev

JSON is an acronym meaning "JavaScript Object Notation". It's a native complex data structure in the JavaScript programming language. It usually represents an "Object"; similar to a C Structure or a Pascal Record.

In a nutshell, it is a programmer-friendly way to structure information in the form of a string. Surely by virtue of its simplicity, JSON has become the format most used to represent information in a transfer between two systems.

Programmers and devops people: Don't assume anything when it comes to JSON. Especially if you aren't in a JavaScript ecosystem. It's not because you have a chain of characters that look like JSON that your favorite programming language will be able to treat it without preliminary treatment.

In JavaScript, you're in luck: the language natively supports it:

let a = {"foo": "bar"};


// bar

But don't expect JSON to work natively in Python:

a = {"foo": "bar"}


# Traceback (most recent call last):
#  File "<stdin>", line 1, in <module>
# AttributeError: 'dict' object has no attribute 'foo'

(I'm hearing pythonista interjecting and talking about Dict, but please hold your peace. We just want to make the point that more often than not JSON strings need to be serialized, deserialized, and require some guesswork depending on the context)

each programming language have built-in functions to serialize, deserialize, load, dump, parse JSON: read the docs and use them

Python does not know JSON natively, we must import the json library to be able to manipulate it in code

import json

json_string = '{"foo": "bar"}'

json_object = json.loads(json_string)


# bar

This json module allows you to deserialize the character string "{'foo': 'bar'}" into a programmatic object. Once deserialized, the programmer will be able to perform different operations on the object. Operations that would be much more difficult to perform directly from a character string!

just be aware that a JSON object and a JSON string are two different things

Most programming languages ​​have specialized functions to serialize and deserialize the JSON. If this is your first time implementing a JSON-based feature in a new system, it's normal for you to struggle a little. Just make sure you have to read the language documentation and make it work by trial and error.

JSON is a way of representing information. It appears most under the form of a character string which must be deserialized into an object to be understood by code. Otherwize your code will only see it as text and fail to parse it.

But, in your daily work, you will realize that certain programs request to receive JSON in the form of a string, while others require you to have it deserialized. Don't assume anything and keep your eyes open to adapt.

For example, this Terraform resource's audit_filter field expects to receive a JSON in the form of a string.

We gladly notice the programmer used the jsonencode utility function to serialize a well-formatted JSON object into a JSON string. It is a fine way to make the code easier to read and leaves it to the compiler to transform into a string when executing the code.

resource "mongodbatlas_auditing" "this" {
  enabled    = true
  project_id =
  audit_authorization_success = true

  audit_filter = jsonencode({
    "$and" : [
        "roles" : {
          "$elemMatch" : {
            "role" : "readWriteAll",
            "db" : "panamaPapers"
        "$or" : [
            "atype" : "authCheck",
            "param.command" : {
              "$in" : [
            "atype" : {
              "$in" : [

this is a API-first world (bonus)

Take for granted any system you work on is either an API or consumes APIs. JSON being the standard payload format, you better make friend with JSON and become fluent in manipulating it.

My suggestion is to never let JSON enter your system's logic. If you want to stay sane, you will deserialize received JSON payloads into your own types and you will serialize your objects into JSON responses just before returning them.

Golang makes it easy and straightforward to do just that by letting you annote each field of a class with json key mapping. You'd be a fool not to use it.

Here is class representing a social media (X, Mastodon) card:

Card       struct {
		AuthorName       string `json:"author_name"`
		AuthorURL        string `json:"author_url"`
		Blurhash         string `json:"blurhash"`
		Description      string `json:"description"`
		EmbedURL         string `json:"embed_url"`
		Height           int    `json:"height"`
		Html             string `json:"html"`
		Image            string `json:"image"`
		ImageDescription string `json:"image_description"`
		Language         string `json:"language"`
		ProviderName     string `json:"provider_name"`
		ProviderURL      string `json:"provider_url"`
		PublishedAt      any    `json:"published_at"`
		Title            string `json:"title"`
		Type             string `json:"type"`
		URL              string `json:"url"`
		Width            int    `json:"width"`