Blog

0001: Documenting Architectural Decisions

Context

Architecture for agile projects has to be described and defined differently. Not all decisions will be made at once, nor will all of them be done when the project begins.

Agile methods are not opposed to documentation, only to valueless documentation. Documents that assist the team itself can have value, but only if they are kept up to date. Large documents are never kept up to date. Small, modular documents have at least a chance at being updated.

Nobody ever reads large documents, either. Most developers have been on at least one project where the specification document was larger (in bytes) than the total source code size. Those documents are too large to open, read, or update. Bite sized pieces are easier for for all stakeholders to consume.

One of the hardest things to track during the life of a project is the motivation behind certain decisions. A new person coming on to a project may be perplexed, baffled, delighted, or infuriated by some past decision. Without understanding the rationale or consequences, this person has only two choices:

  1. Blindly accept the decision. This response may be OK, if the decision is still valid. It may not be good, however, if the context has changed and the decision should really be revisited. If the project accumulates too many decisions accepted without understanding, then the development team becomes afraid to change anything and the project collapses under its own weight.
  2. Blindly change it. Again, this may be OK if the decision needs to be reversed. On the other hand, changing the decision without understanding its motivation or consequences could mean damaging the project’s overall value without realizing it. (E.g., the decision supported a non-functional requirement that hasn’t been tested yet.)

It’s better to avoid either blind acceptance or blind reversal.

Decision

The key words MUST, MUST NOT, REQUIRED, SHALL, SHALL NOT, SHOULD, SHOULD NOT, RECOMMENDED, NOT RECOMMENDED, MAY, and OPTIONAL in this document and all following ADRs are to be interpreted as described in BCP 14 RFC 2119 RFC 8174 when, and only when, they appear in all capitals and with strong emphasis applied, as shown here.

We MUST keep a collection of records for “architecturally significant” decisions: those that affect the structure, non-functional characteristics, dependencies, interfaces, or construction techniques.

An architectural decision record MUST be a short text file in a format similar to an Alexandrian pattern. (Though the decisions themselves are not necessarily patterns, they share the characteristic balancing of forces.) Each record describes a set of forces and a single decision in response to those forces. Note that the decision is the central piece here, so specific forces may appear in multiple ADRs.

We MUST keep ADRs in the project repository under adrs/NNNN-title.md.

We MUST use the GitHub Flavored Markdown dialect of Markdown for text formatting.

ADRs MUST be numbered sequentially and monotonically. Numbers MUST NOT be reused.

If a decision is reversed, we MUST keep the old one around, but mark it as superseded. (It’s still relevant to know that it was the decision, but is no longer the decision.)

Accepted ADR MAY be changed to fix these types of errors1:

  1. Editorial - spelling, grammar, punctuation, or syntax error that does not affect the technical meaning
  2. Technical – error in the technical content (Note that changes in the usage of keywords defined in BCP 14 (RFCs 2119 and 8174) are considered technical.)

We MUST use a format with just a few parts, so each document is easy to digest. The format has just a few parts.

Title These documents have names that are short noun phrases. For example, “ADR 0002: Deployment on Ruby on Rails 3.0.10” or “ADR 0009: LDAP for Multitenant Integration”.

Status A decision MAY be “proposed” if the project stakeholders haven’t decided on it yet, “accepted” once it is agreed, or “rejected” if not. If a later ADR changes or reverses a decision, it may be marked as “deprecated” or “superseded” with a reference to its replacement.

Context This section describes the forces at play, including technological, political, social, and project local. These forces are probably in tension, and should be called out as such. The language in this section is value-neutral. It is simply describing facts.

Decision This section describes our response to these forces. It is stated in full sentences, with active voice. “We MUST …”

Alternatives This section describes considered alternatives to the decision. Each alternative MUST have a Verdict specifying the reason it was not choosen.

Consequences This section describes the resulting context, after applying the decision. All consequences should be listed here, not just the “positive” ones. A particular decision may have positive, negative, and neutral consequences, but all of them affect the team and project in the future.

Resources This section lists references, sources, or further reading recommendations that were used to form the decision or provide an additional context.

The whole document should be one or two pages long. We will write each ADR as if it is a conversation with a future developer. This requires good writing style, with full sentences organized into paragraphs. Bullets are acceptable only for visual style, not as an excuse for writing sentence fragments. (Bullets kill people, even PowerPoint bullets.)

Alternatives

Do not record significant decisions.

Verdict: Not acceptable. Not recording architectural decisions will not be in the spirit of this project.

Consequences

One ADR describes one significant decision for a specific project. It should be something that has an effect on how the rest of the project will run.

The consequences of one ADR are very likely to become the context for subsequent ADRs. This is also similar to Alexander’s idea of a pattern language: the large-scale responses create spaces for the smaller scale to fit into.

Developers and project stakeholders can see the ADRs, even as the team composition changes over time.

The motivation behind previous decisions is visible for everyone, present and future. Nobody is left scratching their heads to understand, “What were they thinking?” and the time to change old decisions will be clear from changes in the project’s context.

Resources

  1. Unlike RFC Errata we allow Editorial and Technical types of errors to be fixed in the ADR itself. ↩︎