.self: A new top-level domain designed to support self-hosting
A proposal for a new `.self` top-level domain aims to give every person a free, human-centered address for self-hosted services, with features like shared mail servers and a “one person, one subdomain” policy to curb squatting. Commenters question whether a dedicated TLD is necessary versus using existing domains, and raise concerns about governance, identity verification, abuse prevention, and the high cost of running a TLD under ICANN’s rules. Supporters like the vision of lowering barriers to homelabs and personal infrastructure, but many doubt its practicality, security implications, and long‑term funding model.
Purpose and Vision of .self
- Proposed as a new gTLD specifically for self‑hosting and small homelabs.
- Goals: one-person-one-subdomain, free personal subdomains, shared mail server, integrated TLS, and treating the TLD as a public good.
- Framed as “human‑centered” and aligned with digital sovereignty / personal identity online.
Why a New TLD vs Existing Options
- Proponents argue owning the TLD avoids dependence on existing registries, lets them build in custom features (mail, certs, tooling) from the root, and keeps names concise.
- Critics say everything could be done today under an existing domain (e.g., .net, .cloud, duckdns‑style) without ICANN complexity and cost.
- Some see the TLD-first approach as “solving the hardest side quest first” instead of shipping working services now.
ICANN, Cost, and Funding Model
- Application fee is ~US$227k; project says it qualified for ICANN’s Applicant Support Program to reduce this but will still pay a significant amount.
- Running a TLD also entails ongoing ICANN fees, policy/compliance overhead, and registry infrastructure (DNS, EPP, RDAP).
- Plan is sponsorships and donations, likened to Let’s Encrypt/ISRG; skeptics note those had heavyweight institutional backers and very concrete technical plans from day one.
One Person, One Free Domain – Feasibility and Abuse
- Ambitious promise: every person gets one free subdomain, no parking, squatting, or reselling.
- Multiple commenters question enforcement:
- How to reliably prove “one human” without invasive ID checks.
- Risk of people in poor regions reselling their free domains to spammers.
- Heartbeat / activity checks are easier for squatters to automate than for casual users.
- Ideas floated: ID-based naming, zero‑knowledge identity systems, proof‑of‑use schemes, lotteries, bans on transfers, but all seen as complex or exclusionary. Overall feasibility is “unclear.”
Spam, Security, and Reputation Concerns
- Widespread fear that a free TLD will repeat the .tk story: mass abuse, phishing, eventual blanket blocking by large providers.
- Shared mail server plan prompts skepticism about getting Gmail/Outlook to trust .self mail; some see central mail as anti‑self‑hosting or creating a single failure/abuse point.
- Several worry a “self‑hosted” brand might advertise weaker targets to attackers.
Namespace Bloat and Governance
- Some argue new gTLDs “pollute the commons” and add user confusion with little public value versus .me, .name, etc.
- Others criticize ICANN as a money grab / quasi‑cartel and point to US jurisdiction over almost all non‑ccTLDs.
- Governance, dispute handling, and long‑term commitment to free domains are seen as hard, under‑specified problems.
Alternatives and Practical Self‑Hosting Today
- Many note existing options: cheap TLDs like .xyz, ccTLD hacks (.me, .ai, .io), and internal domains (.home.arpa, .internal, .lan, mDNS/.local).
- Some say the real issue is making self‑hosting easier (DNS tooling, dynamic IP handling, tunnels, OS distros) rather than inventing another TLD.