Information technology — SoftWare Hash IDentifier (SWHID) Specification V1.2
Introduction: Modern software relies heavily on open source components that are developed collaboratively in a distributed setting, and that are assembled to create complex systems that evolve at a fast pace. This has strengthened the need to precisely track, ensure availability, and guarantee integrity of the components that go into a given system for a variety of stakeholders. Academia needs to ensure that research results are reproducible, industry needs to improve the traceability of the software supply chain, and developer communities need tools to cope with the increasing complexity. A key building block for addressing this issue is a system of intrinsic identifiers that allows users to precisely pinpoint the exact version of any software artifact, at all levels of granularity, without relying on any central registry or naming authority. With this specification, the SWHID working group makes such a system of intrinsic identifiers, originally developed for the Software Heritage universal source code archive,[1] available to all stakeholders. For the sake of clarity, examples have been drawn directly from the Software Heritage archive; however, it is important to note that systems for the persistent archival of software artifacts, as well as resolution of SWHIDs, are outside the scope of this specification, which does not require the use of Software Heritage. © ISO/IEC 2025 – All rights reserved
Scope: This specification defines a standard data format for referencing software artifacts that match the data model of modern distributed version control systems. This format includes the typical tree-like structure of a filesystem hierarchy, but also, special nodes to track revisions and releases, as well as the full status of a version control system, with all its development branches. A key property of SWHIDs is that they can be computed using cryptographically strong functions directly from the digital objects they refer to, by anyone that has access to a copy of those objects. This enables decentralised and independent verification of integrity, without relying on a registry or a central authority. The computation of the SWHID identifiers is based on Merkle Acyclic Directed Graphs, a natural generalization of Merkle trees. The resolution of SWHIDs, that is, the process of obtaining a copy of a digital artifact corresponding to a given SWHID, is outside the scope of this specification.