At Airbus, documentation (100 separate manuals) for the 380 is handled by 1,000 people around the world, each one working on different subsystems. This information is delivered in applications and formats that are appropriate to the jobs of the target audiences (pilots, mechanics, flight engineers, baggage handlers, etc.).

What can we learn from the best practices that Axel Sellmer and his documentation team have applied in designing and deploying Airbus’s new distributed documentation applications and workflows?

• They begin with granular information objects that can be combined and recombined.
  
  
• They use a layered information architecture—where each person worries about the stuff in their layer.
  
  
• Information assembly is kept distinct from information creation and maintenance.
  
  
• The data and the information about the data are kept separate.
  
  
• Each information object can be linked to multiple data objects.
  
  
• Each information object has a well-defined lifecycle.
  
  
• Change is constant.
  
  
• Visibility and transparency are built in. Anyone can see the current status of any information object and its interdependencies.
  
  
• The use of standards-compliant metadata is crucial. Each plane’s information collection needs to last 50 years.

For those of you with less elaborate information and documentation needs, I believe that it makes sense to start with many of the same principles used in designing the information architectures for these complex requirements and then “skinny it down.” While these principles apply well to highly-structured and -regulated information environments, the fact that each information object is self-contained and loosely-coupled to other (data, media, etc.) objects that are also changing over time, makes this a good architecture for organically-evolving ecosystems.