A practice is the actual doing of work — activity conducted according to method. Diagnosing a patient, filing a report, reviewing a pull request, standing a watch, conducting an interview — these are practices. They have rules (written or implicit), roles (who does what), settings (where they happen), and products (what they generate).

Three properties make something a practice:

Activity. A practice is something done, not something described. Practice is action, not documentation. The method describes how to do it; the practice is the doing. You can read about surgery without practicing surgery. You can write a deployment checklist without deploying.

Accordance. A practice follows (or departs from) a method. Practice without method is improvisation — it may work but it is not reproducible. Method without practice is shelfware — it exists on paper but governs nothing. The productive relationship is mutual: method constrains practice; practice tests method.

Situatedness. A practice happens in a specific context — this session, this patient, this file, this decision. Practice is always concrete. A method is abstract (it describes how things should be done in general). A practice is particular (it is how things are being done right now, here, by this person, under these conditions).

Practices produce knowledge. A clinical practice produces case files, diagnoses, and statistical distributions. A research practice produces observations, data, and publications. A coding practice produces code, tests, and documentation. This knowledge is not a neutral byproduct — it is shaped by the practice that produced it. Different practices on the same subject produce different knowledge.

Foucault traced how practices change rather than how ideas change: the prison’s timetable came from the monastery, its drill from the military, its examination from the school. Practices carry their history. The job interview, the clinical examination, the standardized test — none are inevitable. Each emerged from specific conditions and could have been otherwise. Practices are contingent.

Bourdieu’s habitus is the internalized disposition that generates practices without explicit rule-following — the skilled practitioner who acts correctly without consulting the method. Dewey’s pragmatism treats practice as inquiry — doing is a form of investigation, and the results of practice feed back into the methods that govern it.

Practice is to method as performance is to score, as speech is to grammar, as execution is to plan. The relationship is constitutive in both directions: method without practice is inert; practice without method is unaccountable.