It's funny how many organisations insist on trying to control the informal through formal methods.

What do I mean by that?

Well there is this stubborn idea in (often larger) organisations that we should control the collective knowledge. You know the activities we perform when we are asked to document it on a wiki page, or write down how you work into a process, or my favorite .. being asked to "hand over" what you know to your colleague when leaving a work place.

To anyone who knows anything about tacit vs. explicit knowledge these attempts always seems rather foolish, wasteful and in the later instance way to late.
Don't get me wrong there are things that we can, and should, document (i.e move from brain onto paper).
It could be "how do we build this system" or "what are the dependencies for component X".
Straightforward questions that are more or less easy to answer.

"We can know more than we can tell" - M.Polanyi

But when we are asked to document something like "How do you fix a bug", or "How do you ride a bicycle", then things get complicated.

This type of knowledge can at best be "captured" when the person holding that information joins a community of practice or similar.
In that case the knowledge would be transferred directly in between individuals and through direct dialogue and interactions.

Since tacit knowledge was defined in 1958 it's been well known in the field of knowledge management that some knowledge is very hard to "transmit" to another person via words and symbols.
In fact some academics don't even want to use the word transmit in this context as it implies that knowledge can be packaged and sent, they rather talk about learning and socialization instead.
And yet some (far too many) managers and organisations insist on this practice of trying to formalize what is informal.

In the past I've worked as a consultant for a "large telecommunications equipment" company.
In that organisation there has been several attempts, at times together with the academia, to create catalogues/index:es/"phone books" where the purpose has been to document what knowledge exists in the organisation, who has that knowledge, or which person knows some other person that you can ask for help in a specific area.

The first problem with such an attempt is that at best (and I'm being generous here) the second you are "done" with such an attempt, then that information is already outdated (people quit, move to other departments, technology changes, and so on).
But there is also the aspect that the knowledge we are after is most of the times more dynamic and most of the times we can't even express exactly what we want answered.
Just trying to describe how you would go about "google:ing" for a specific question is challenging and depending on which person you ask you would probably end up with different answers.

Another classic for us "computer literates" within the family is when we are asked to fix a computer, or solve some problem. Afterwards the relatives sometimes asks us how we did it, and I for one is most of the time having a hard time explaining what I did.
It normally involves exploring the menus looking for things that might match, searching a bit online, clicking a bit more and by some combination of experience and intuition ending up solving the problem at hand.

I suspect that one of the reasons why organisations resort to these fruitless attempts is that there is a fear of being dependent on an individual. The reasoning going that if they leave they take the knowledge with them.

But with all the de facto results showing that this is a waste of time and resources why don't managers try something else?
Being a bit blunt I think one of the reasons for this is simple incompetence... to many managers are in their position without having too much insight into people, building organisations, and knowledge management.
It's a sad fact, but to many managers out there should simply put not be managers at all, luckily I've never worked directly for one, although I've encountered way to many of them when I was a consultant.

Another reason is that a small number of academics out there that stubbornly try to come up with ideas that "this time" should solve the problem, and comes up with fancy terms like "knowledge exchange protocol" (shivers).

And some management consultants also sells these ideas when talking to management.

But if managers shouldn't be afraid of being dependent on their co-workers, what should they do instead?
Why not trying to actually encourage building informal networks, communities of practice, and helping your co-workers building one-to-one relations within your organisation.
Encourage people working together in pairs or teams, and just advocate face-to-face meetings and dialogues.

You don't learn how to ride a bike from a wiki-page or a document describing the process.
It's done when one individual talks and explains it directly to another person, and then that person trying to ride the bike.

Because what's being done today is clearly not working (can anyone say "outdated wiki pages"), so lets try something else instead.
Put trust in the people you manage and help them, the organisations they work in grow and increase the collective knowledge instead.
People might actually enjoy that more and stay even longer with you.