The worst problem I have had with the corpus of documents comes from my experience bringing on new people at NCSA. People cannot pick up a design document and understand 1) what is invented by the author of the document v.s 2) what in the document is reflective of requirements or other constraints in governing doc sets.
Or in a phrase, there is no traceability.
I am working on making changes to LDM-230 so we can make a reviewable earned value plan for the L1 system. I’m currently working in Confluence so I can take the input of others and be in an environment where linking is possible, and might be of use implementing traceability.
I really think this is a first-class problem in the project, and I’s hope it could be addressed.