Designing the Time Machine together

Requests for Comments

Motivation

Reaching consensus on the technology options to pursue in a programme as large as Time Machine is a complex issue. To ensure the open development and evaluation of work, a process inspired by the Request for Comments (RFC) that was used for the development of the Internet protocol IETF RFC 791 is being adapted to the needs of Time Machine. Time Machine Requests for Comments are freely accessible publications, identified with a unique ID, that constitute the main process for establishing rules, recommendations and core architectural choices for Time Machine components.

Approach

  1. Accessibility. RFCs are freely accessible, at no cost.
  2. Openness. Anybody can write an RFC.
  3. Identification. Each RFC, once published, has a unique ID and version number. It can nevertheless be revised over time as a living document, being republished with the same ID and a different version number.
  4. Incrementalism. Each RFC should be useful in its own right and act as a building block for others. Each RFC must be intended as a contribution, extension or revision of the Time Machine Infrastructure.
  5. Standardisation. RFCs should aim to make use of standardised terms to improve the clarity level of its recommendation.
  6. Scope. RFCs are designed contributions and implementation solutions for solving practical problems. RFCs are not research papers and may not necessarily contain experimental evidence. RFCs cover not only the technical infrastructure but the data standards, legal frameworks, and values and principles of Time Machine.
  7. Self-defining process. As used for the development of the Internet, RFCs are the main process for establishing Time Machine Infrastructure and Processes and also the processes and roles for managing RFCs themselves.

Publication process

The RFC Editorial Committee organises the publication process of the RFCs, maintains the consistency of the RFC System, appoints RFC teams to organise new RFCs and to improve existing RFCs, keeps track of RFC versioning, ensures the timely and regular publication of RFCs, and is responsible for the public announcement of the open review process. The governance and organisation of the RFC Editorial Committee is defined in RFC-0004.

The publication process is the following :

  1. The RFC Editorial Committee appoints authors to write the RFCs planned in the RFC tree (RFC-0002). Alternatively, authors may contact the RFC Editorial Committee to submit their candidature to write an RFC (planned in the RFC tree or not).
  2. The authors produce an RFC draft which is reviewed, first by the RFC Editorial Committee for coherence with the rest of the RFC corpus and then by a larger community. The RFC is revised and possibly sent for review again.
  3. Once accepted by the RFC Editorial Committee, an RFC receives an official identifier and is officially published as an peer-reviewed publication with proper scholarly credits assigned to the original author(s).
  4. The RFC tree is adapted to include the published RFC and any possible sub-RFCs planned during the writing of the RFC.

RFC Repository

The RFC draft and publication documents are hosted on the RFC Repository on GitHub. Have a look and join us there with comments and text contributions.