Aside from purely functional protocols that operate at a particular layer in the stack and are primarily used by the machines themselves, some protocols are created for humans to interact either with the system or through the system to another human user. Both types of protocols rely on assertions, however human interface protocols tend to rely on them more heavily which gives them a distinct character. These assertions are actively governed by human institutions, and since human interface protocols rely on assertions more heavily than functional ones, it is more likely for human idiosyncrasies to be reflected in the technical implementation of a human interface protocol.
DNS is one of these human interface protocols. At its core, DNS maps a character string to an IP address, which is the protocol computers use to find each other on a network. There are some purely technical reasons for using a domain name, which remains constant, rather than an IP address, which can change dynamically, but the DNS naming convention was deliberately designed to be human-readable and memorable. All domain names are nested within a hierarchical structure, at the top of which is the root zone. At the first level below the root zone lie the Top-Level Domains (TLDs) like .com, .org, and others. The root zone, all the TLDs, and every subdomain beneath them is an assertion set by one or many humans. The root zone, being the foundational assumption that the entire protocol is based on, is managed by the Internet Corporation for Assigned Names and Numbers (ICANN), a massive international NGO which establishes the rules, procedures, and protocols for everything that has to do with DNS globally. They decide what the TLDs are, who operates them, and who is allowed to create new entries in them, among other things.
As of this writing, there are 1,593 TLDs defined in the root zone and ICANN has just opened only its fourth new gTLD application round since its founding in 1998, a rigorous process that takes years to complete. ICANN has such a measured approach because the stakes are so high: if anything were to break in the root zone, the entire Internet would cease to function until it were fixed. Ultimately, what ICANN exists to do is to carefully curate and vet the assertions that define the core of the system before putting them into place.
If ICANN specifically did not exist, then there would still be some other person or entity which occupies that position. Indeed, as the Internet was in its most nascent phase and before ICANN existed, it was effectively “just some guy” who held this position of absolute authority and control over DNS simply as a matter of emerging circumstance. His name was Jon Postel and he played the role of ICANN before ICANN existed. Postel was hugely influential over the creation of the Internet in general but the focus here is that this position, first held by Postel and now by ICANN, is inseparable from the protocol itself and can effectively be considered as part of it. The position has nothing to do with ones and zeros being carried by electrical signals, and yet those ones and zeros are shaped entirely by human assertions which define how everyone interacts with the Internet. The average person does not know anything about the technical specification for the ones and zeros implementing the DNS protocol and has never heard of ICANN, they just know to type “something dot com” into a browser to connect to a website. To them, the reasons for the technical specification are indistinguishable from the reasons that the website’s URL ends in .com and not some other random set of letters.
The key difference between a purely functional protocol and a more human interface protocol is that the assertions in a functional protocol become fixed upon creation whereas the assertions in human interface protocols can be changed over time. Rather than change a functional protocol after it has been defined, it is more practical for a new version to be created, whereas a human interface one can adapt by changing the assertions. The voltage and frequency expected of a given signal protocol will never change but the number and type of TLDs that exist will. For human interface protocols, power lies in controlling their assertions