]> Bluesky Social

bnewbold@robocracy.org

Applications and Real-Time Area Authenticated Transfer This document defines the "at" identifier scheme, which is used to reference accounts and data records in the Authenticated Transfer Protocol. About This Document Status information for this document may be found at . Discussion of this document takes place on the Authenticated Transfer Working Group mailing list (), which is archived at . Subscribe at . Source for this draft and an issue tracker can be found at .

Introduction The Authenticated Transfer Protocol enables the creation of decentralized networks for publication of self-certifying data. An introduction to the overall protocol architecture is given in , and the data repository and synchronization mechanisms are described in . Each account has a global permanent account identifier that can be resolved to a network hosting location and to public key material. Multiple account identifier systems are supported, but details are out of scope for this document. Accounts publish structured data records of different application-defined types in data repositories. Each account has a single repository for all of its public data records, organized in collections by record type, with one or more records in each collection. Individual records can be globally referenced by account, collection, and record key. Records themselves may include references to other records, forming a global data graph. Record references can be resolved to fetch individual records. They can also be used when annotating records for the purpose of content moderation. This document describes a string identifier syntax for referencing data records and entire accounts. The identifier syntax described in this version of the document does not comply with the generic Uniform Resource Identifier (URI) requirements. Using account identifiers with multiple colons in the authority section violates the generic syntax rules for URI schemes using the double-slash prefix (//), which this version uses. This means the identifier syntax described in this document is not eligible for permanent registration in the IANA URI Registry under .

Structure The generic structure of an "at" identifier is: The required authority section references an account. It may be a permanent account identifier or an account handle. An identifier that only includes the authority section can be used as a reference to an overall account. Handles in the authority section are discouraged in most other use cases; see . Except for the authority syntax noted in , the structure aligns with the generic structure and semantics described in Section 3 of . An empty authority section is not allowed. Userinfo is not supported in the authority section, and host/port separation with a colon character is not used. The query and fragment sections have no defined semantics and are reserved for future use. The path section can be used to reference a specific resource controlled by the account authority. A common use case is to reference an individual data record from the account's public data repository: The collection part indicates the data record type (schema), and the record key identifies the individual record. The path section matches the path under which records are stored in the repository data structure described in .

Syntax The overall "at" identifier encoded string length limit is 8192 ASCII characters. They MUST NOT include a trailing slash. The authority section can contain either a permanent account identifier or an account handle. The syntax of specific account identifier systems is out of scope for this document, but a few generic syntax restrictions apply to all such identifiers:

• the account identifier string is ASCII, containing letters (A-Z, a-z), digits (0-9), period (.), hyphen (-), underscore (_), and colon (:)
• other ASCII characters may be represented with percent encoding (percent character % followed by two hexadecimal characters)
• must not end in a colon (:)
• length (including percent encoding) is between 1 and 2048 characters
• MUST NOT be a simple account handle

The account handle system is out of scope for this document, but the handle syntax is as follows:

• at most 253 ASCII characters in total
• consists of multiple segments separated by periods (.)
• empty segments and leading or trailing periods are not allowed
• there must be at least two segments, and thus at least one period ("bare" top-level domains are not allowed)
• each segment must have between 1 and 63 characters, consisting of lower-case letters (a-z), digits (0-9), and hyphens (-)
• segments cannot begin or end with hyphens
• the last segment must not start with a digit

Handles MUST be normalized to lower-case when included in "at" identifiers.

Record References An "at" identifier referencing a record has additional syntax restrictions. It has exactly two path segments, and must not include query parameters. The first path component must be a valid Namespace Identifier (NSID) string, as defined in Appendix D of . A non-normative summary of that syntax is:

• at most 317 ASCII characters in length
• overall NSID is case-sensitive
• "domain authority" part (reverse-order DNS hostname, with at least two segments separated by periods) separated from a final "name" part by a period (.)
• domain authority segments are each between 1 and 63 characters; consist of ASCII lower-case letters (a-z), digits (0-9), and hyphens (-); must not start or end with a hyphen; the first segment must not start with a digit
• the final name part is between 1 and 63 characters; consists of ASCII alphanumerics (A-Z, a-z, 0-9); must not start with a digit

The second path component is a Record Key string, with syntax defined in Section 3.1 of . A non-normative summary of that syntax is:

• length between 1 and 512 ASCII characters
• consists of alphanumerics (A-Z, a-z, 0-9), period (.), hyphen (-), underscore (_), colon (:), and tilde (~)
• case-sensitive
• literal values . and .. are forbidden

Examples The following are valid "at" identifiers referencing accounts: The following are invalid "at" identifiers under the syntax defined in this document: The following are valid "at" identifiers referencing a record: The following are invalid record references (though they may be valid generic "at" identifiers):

Security Considerations Record references that use account handles instead of permanent account identifiers can have the authority of the reference change over time. Such references should be resolved to a permanent account identifier before being persisted to long-term storage. References stored in record data should always use permanent account identifiers.

IANA Considerations

URI Scheme Registration As noted in , the identifier syntax described in this version of the document is not eligible for permanent registration in the IANA URI Registry under . If it were, registration metadata would be included in this section.

References Normative References A Uniform Resource Identifier (URI) is a compact sequence of characters that identifies an abstract or physical resource. This specification defines the generic URI syntax and a process for resolving URI references that might be in relative form, along with guidelines and security considerations for the use of URIs on the Internet. The URI syntax defines a grammar that is a superset of all valid URIs, allowing an implementation to parse the common components of a URI reference without knowing the scheme-specific requirements of every possible identifier. This specification does not define a generative grammar for URIs; that task is performed by the individual specifications of each URI scheme. [STANDARDS-TRACK] This document updates the guidelines and recommendations, as well as the IANA registration processes, for the definition of Uniform Resource Identifier (URI) schemes. It obsoletes RFC 4395. Bluesky Social Bluesky Social Informative References Bluesky Social Bluesky Social

Acknowledgments This document is based on the original Authenticated Transfer design work by Paul Frazee and Daniel Holmgren, as described in .