Domain Name
A domain name is a human-readable name, such as example.org, that identifies a realm of administrative control on the Internet and can be mapped to numeric addresses and other data. It is the thing people type, print on business cards and put in email addresses, and it exists so that nobody has to remember a string of digits to reach a service.
A domain name is read from right to left as a hierarchy, with each dot-separated label narrowing the scope. In www.docs.example.co.uk the rightmost label, uk, is a top-level domain (TLD); co is a second-level label beneath it; example is the name a registrant chose; and docs and www are subdomains the owner created freely. Behind all of this is an unwritten empty label, the root, which is why a fully qualified domain name (FQDN) is technically written with a trailing dot. Each label may be up to 63 characters and a whole name up to 253 in text form. TLDs fall into categories: generic ones such as .com, .org and .net, country-code ones such as .uk or .jp, and a large set of newer generic names added in the 2010s.
Domains are leased, not bought. The registry for each TLD maintains the authoritative list of names beneath it, while registrars are the companies that sell registrations to the public, usually for one to ten years at a time. ICANN, created in 1998, coordinates this system, including the root zone and the accreditation of registrars. A registrant who stops renewing loses the name, and after a grace period anyone may register it, a real hazard for forgotten domains that still appear in old links and email. The first .com name ever registered was symbolics.com, on 15 March 1985, a reminder of how early the scheme was settled.
A domain name is not a website, a URL or an address. The name only identifies a location in the naming hierarchy; whether it hosts a website, an email service, both, or nothing at all is decided by records the owner publishes. A URL such as https://shop.example.com/cart?id=5 contains a domain name as just one component, alongside the scheme, path and query. The "www" prefix is merely a conventional subdomain, not a requirement. And the number an application ultimately connects to is an IP address, which the domain name is translated into by the separate Domain Name System rather than by the name itself.
Names can carry international characters through internationalized domain names, which are encoded in an ASCII form beginning with xn-- called Punycode so that existing infrastructure keeps working. That flexibility creates a security risk: a name mixing look-alike characters from different scripts can imitate a familiar brand, an attack known as a homograph attack, and so browsers apply display rules to expose suspicious names. Typosquatting, registering names one keystroke from a popular one, and expired-domain takeover are related abuses. Ownership details historically published through WHOIS are now frequently redacted for privacy.
For programmers and data practitioners, domain names show up everywhere: as keys for cookies and the same-origin policy, as the unit for email authentication records, and as identifiers in logs and scraping pipelines. Extracting the registrable part of a hostname is a classic pitfall, since splitting on dots fails for names such as example.co.uk, where the registrable name is three labels long; correct code uses the Public Suffix List, a community-maintained catalogue of where registrations actually begin. Wildcard certificates, CDN configuration and cloud services all key off domain names too.
Read a domain name as a leased, hierarchical label rather than an address: a memorable name that someone registered beneath a TLD, whose meaning is entirely determined by the records its owner chooses to attach to it.