Documentation is not intended to contain configuration. Configuration is meant to be consumed by scripts or similar tools.

As code, configuration should be readable. As code, effective (production) configuration is the reference. Documentation may be generated from it, but not the opposite.

As an example, consider system architecture. In some “Architecture” document, you describe relationships between servers and clients:

  • you may put some schemas or diagrams in the documentation, so that you give users an overview of the architecture.
  • you shouldn’t write server names in the documentation. You’d better reference the configuration of your environments.
  • utilities you use for deployment, network management or monitoring should provide comprehensive views, or at least entry points so that you could generate comprehensive views.


Here is an overview of the architecture:

.. image:: /_static/architecture-diagram.svg

Architecture configuration is consumed by deployment tools:

* `Production <>`_
* `Staging <>`_

Monitoring gives you information about servers:

* `Production <>`_
* `Staging <>`_


In production environment:

* ```` is a Debian Squeeze server with, 4Go RAM and 20Go HDD.
  It serves:

  * the frontend, with Django 1.2
  * PostgreSQL server, version 8.4
  * Memcache

* ```` is a FreeBSD server with 256Mo RAM and 500Go HDD.
  It serves static files with Nginx.

Configuration will not be maintained in documentation. Thus it is to become obsolete, wrong and could lead to errors, misunderstanding... i.e. it has negative impact.