There's an old joke that us IT folks find very funny.

Q: Why could God create the world in only seven days?
A: There was no installed base

I'm deep in the process of implementing a corporate content management system. The software base is Confluence, a wiki with a lot of enterprise friendly tools.  The test system is in place and I've started the process of seeding it with documents. 

So far, I've identified four primary sources of information.  There's a big, shared folder called "S-Data". Within that folder, there are folders named by functionality and individuals.  The second information source is personal folders. Each person also has their own personal folders on the big file server.  Sometimes important corporate documents are in the S-Data folders, sometimes in their personal folders. Email (specifically, the Exchange server) is the third information repository.  There are things recorded in email that are recorded no where else.  In addition, the Exchange server hosts public folders that function as a CRM and document exchange.  Finally, there's an intranet hosting documents and discussions, with contributions from the corporate office and a distributed network of affiliates.

It has struck me that I'm not implementing a content management system.  I'm implementing a Knowledge Base, a system that should contain the sum of knowledge about the current state of the organization. Going forward, the version control features of Confluence will create a historical record, documenting what has changed over time. Strategically, this is A Good Thing.

Strategically, yes.  But, oh my, the task of organizing and sifting is proving painful.  It would be so much easier to buy a Google Search Appliance, let it index all the content, and put up a search page and call it "Our Knowledge".  You and I both know that's a straw man argument.  It finds everything, but still doesn't give you history, context, or control.  Using a search tool Band-Aids the lack of consistent structure, framework, and planning. It's a short term fix that lets us avoid the painful process of changing thought and business processes.

It's time to think hard about how the knowledge trees are structured.  The goal of the project is a Wikipedia of corporate knowledge, deeply crosslinked, with the most commonly accessed information most closely at hand.  I don't yet have the right tree structure, but I have one overriding organizing principle:  No Band-Aids.