Marc Stiegler — Petnames and Introduction Patterns

<!-- hierarchy:nav -->

Path: OIP › Thinker Reference › Thinkers › Marc Stiegler — Petnames and Introduction Patterns

Shelf: Thinkers · Traversal: self-explaining · hierarchical · voxel-ready
Machine root: OIP tree · Registry

Marc Stiegler — Petnames and Introduction Patterns

§SELF — thinker-marc-stiegler

What this page is: A profile of Marc Stiegler and his work on petnames and introduction patterns in capability security. What it explains: What petnames are, why they replace global naming systems, and how introduction patterns let two parties meet safely through a mutual contact. Why read it: To understand how local naming solves the conflict between human-readable names and decentralized security, and how OIP's plain-language resolution (?ask=) applies the same principle.

What Petnames Are

A petname is a local name that a user assigns to a capability (a cryptographic token that grants access to a resource). The user chooses the name. No central registry is involved. Two different users can assign different petnames to the same capability without conflict.

The Problem Stiegler Solved

Cryptographic capability tokens are long random strings. Humans cannot remember or type them. Global naming systems like DNS or usernames attempt to solve this, but they require central authorities to resolve conflicts (two people cannot both own "alice") and they create targets for attack (control the registry, control the names). Stiegler asked: can humans use readable names without a central authority? His answer: yes, if names are local.

The Key Idea

If you receive a capability from someone, you name it yourself. You might call it "Cyrus's Token." The person who gave it to you might call it "Owner Key." The same underlying capability has two petnames — one in your namespace, one in theirs. There is no global name. There is no collision. There is no authority to appeal to or attack. The name exists only in the relationship between you and the capability.

Introduction Patterns

Stiegler also designed introduction patterns: protocols for safely introducing two parties who do not know each other. The pattern works through a mutual introducer who vouches for both parties. The introducer does not become a communication bottleneck or a point of trust beyond the introduction itself. Each party receives a capability from the introducer that lets them communicate directly with the other party. After introduction, the introducer is no longer needed.

What Stiegler Got Right

• Petnames eliminate the naming authority problem entirely. No central registry means no central point of failure or control.
• Local naming matches how humans actually name things. People already use different names for the same person ("Mom," "Dr. Chen," "Jennifer").
• Introduction patterns distribute trust without concentrating it in intermediaries.
• The separation between capability (the access token) and petname (the human label) is clean and orthogonal.

What Stiegler Got Wrong or Left Unfinished

• Petnames do not solve the initial discovery problem. You still need some channel to receive a capability before you can name it. The petname system starts after the first contact.
• There is no standard data format for exchanging petnames between systems. Each implementation invents its own.
• Introduction patterns assume honest introducers. A malicious introducer can misrepresent one party to the other during introduction.
• The model does not specify revocation. If a capability is revoked, the petname may still point to a dead or repurposed token.

How It Connects to Other Ideas

• Zooko's Triangle: Petnames are the standard solution to Zooko's Triangle, which states that a name cannot be global, secure, and human-meaningful all at once. Petnames sacrifice globalness to achieve security and human-meaning simultaneously.
• Object Identity Protocol (OIP): OIP's ?ask= resolution is a petname system. The user supplies words in their own vocabulary; the system resolves those words to the correct object capability. The user chooses the words; the system performs the mapping.
• Capability Security: Petnames are a usability layer on top of capability-based access control. They do not replace capabilities; they make them usable by humans.

Sources

• Stiegler, Marc. "An Introduction to Petname Systems." 2005.
• Stiegler, Marc and Mark Miller. "The Structure of Authority: Why Security Is Not a Separable Concern." HP Labs.
• Miller, Mark S., Marc Stiegler, and Bill Tulloh. "The Digital Path" (work on capability patterns at Combex and HP Labs).

---

Up the tree

• OIP root — protocol root, zero-context entry
• Thinker Reference hub — full hierarchy map
• Thinkers shelf — siblings on this shelf
• Voxel graph article — how pages link as voxels
• Self-describing protocol

Related on this shelf

• Alan Kay — The Big Idea Is Messaging
• Alfred North Whitehead — Process and Reality
• J.L. Austin and John Searle — Speech Acts
• Barbara Liskov — Abstract Data Types and Distributed Consensus
• Bram Cohen — BitTorrent and Content-Addressed Protocol Design
• Butler Lampson — Protection and Access Control
• Carl Hewitt — The Actor Model
• Charles Sanders Peirce — Signs, Abduction, and Pragmatism

Machine surfaces

• Public page: https://miscsubjects.com/a/thinker-marc-stiegler
• JSON article: https://miscsubjects.com/api/articles/thinker-marc-stiegler
• OIP ask: https://miscsubjects.com/api/dispatch?ask=Marc%20Stiegler%20%E2%80%94%20Petnames%20and%20Introduction%20Patterns