`NULL` isn’t falsy

Furthermore:
NULL isn’t truthy, unknown, undefined, nonexistent, nil, or zero

Some languages are misleading coders about whatNULL means. A recent project highlighted just how tricky this can be. For example in JavaScript:

let maybeNull = null;
let isNull = maybeNull === null;
let isntNull = maybeNull != null;
console.log({isNull, isntNull}); // {isNull:true, isntNull:false}

JavaScript seems to provide the intuitive answer at first glance. However, as we move forward, you’ll see that this could be misleading. For example, once you connect to a Postgres database:

Do not write expression = NULL because NULL is not “equal to” NULL. (The null value represents an unknown value, and it is not known whether two unknown values are equal.)
— Postgres Docs

Therefore, attempts to compare NULL in Postgres always return NULL(which your JavaScript code may interpret as always false). Since the purpose of a database is to record known information, it seems reasonable that Postgres treats NULL as unknown. Yet this still doesn’t work as a universal definition. For example in GraphQL:

{
  "data": { "myData":NULL },
  "errors": [ ... ]
}

The myData property returned as NULL. This doesn’t mean that myData is unknown to the server. Perhaps GraphQL returned NULL due to authentication or throttle error(s). The value may be known and explicitly defined; you simply need to modify/retry your query.

You might be tempted to test myData === null but this seems like an anti-pattern. Depending on your programming language, the test may always yield false! Furthermore, this doesn’t tell you whether you need to modify/retry your query. Better to resolve the errors and parse the expected type first (nicely done with zod). If you resolved all errors and didn’t obtain an expected type, it is de-factoNULL (and therefore no need to test it).

Is the dictionary any help?

Some definitions for NULL are helpful:

Void; of no legal or binding force or validity; of no efficacy; invalid.
Of no account or significance; having no character or expression; negative.
Something that has no force or meaning; that which is of a negative or meaningless character.

However, with regard to programming, some definitions of NULL can also be misleading:

Not any; wanting; non-existent.
Zero in algebraic measure.

So how should we define it?

My opinion, is something along these lines: NULL is similar to “I plead the 5th” or an interview non-response due to a non-disclosure agreement. The value will be unknown to you; however this does not mean that it is unknown to the discloser. It may be anything, everything, or nothing. It may be known or unknown.

Thinking through database design

Suppose the following Postgres table:

| User    | URL                          |
| ------- | ---------------------------- |
| Bob     | https://example.com/bob      |
| Betty   | https://example.com/betty    |

We create new user Pat before obtaining a URL. Should URL be NULL or something else? I’d argue that the default should be UNKNOWN if possible.

• NULL doesn’t provide the most helpful information
• empty string is not a valid url
• VOID suggests the URL is known (and nonexistent)
• UNKNOWN explicitly confirms that this value is not currently known

Now suppose Betty removes her URL (now known to be nonexistent). It seems VOID would be most appropriate.