But it’s not only the documents creators who should invest efforts, readers must invest enough to be familiar with organization and knowledge domain. Even when search functionality works great, you would still need to know what to search for.

Thanks again for the feedback.

I would also add that writing the documentation is only useful if you ensure it’s properly distributed or can be easily found. Finding the correct location for the documentation and ensuring that it’s organised in such a way to make retrieval easy is not as simple as it could be in a large company(at least that’s what I’m observing where I work atm).

After just reading about BAAG for the first time and the motiviation behind it, I can agree with and understand the need for battling this.
And yet I am not convinced by the attack on guessing and the need to eliminate it.
The need (as put on your about page) to eliminate guesswork from the decision making process is to deny the basics for how the human thought process works.
The brain based on neural networks, does not use logic for processing information. To ultimately try and remove a core part of the human thinking patterns where we
have to make constant assumptions all the time to avoid information overload is a natural thinking strategy. It also allows us to think outside the box and not become mindless automatons.
I would see a better solution is to not battle it, but to train people in the correct use of guessing versus logical deduction and when each method is best. The examples you provide highlight the need for educating people more on this, rather than battling against it. By alienating guessing, by making it appear bad you are deliberately driving home a point for a reason, which I can understand. After all, I have done this myself, and its a good tactic - only when used right.

When you move into the multimedia worldview and store it in the database, logic that we are used to disappears and there is an element of calculated guessing all the time.
For example, compare two photos and ask which is the better one? It cannot be logically calculated. And sometimes when it comes to performance tuning and problem resolving in a mission critical system that has time constraints, guesswork is crucial to quickly resolve it (I will not give any Star Trek analogies here especially Movie #4, but will try and track down some Hitch Hiker Guide to the Galaxy ones). Though it is guesswork based on experience and knowledge.

I believe with the passion you have raised on this topic, it is worth discussing it further. I am now aware of this issue, which I wasn’t before this morning began.

Marcel.

To help remedy that, here’s my recent post on system administrator site documentation with docbook, and a few “intro to docbook” links that were helpful to me.

I upgrade the sql2005 stuff until a very recent release, something SP2′ish.
For, eh, legal/corporate(employer/business guest at MSFT) and so reasons I stopped this hacking work.

In SQLServer 2008 there is still no dmv (v$ for Oracleites) view per session or so.
But there is something really new and interesting that can present wait events: XEvents.
With XEvents you can trace every wait ( and a lot of other events) and get extra context information when they happen.
I presented on this subject on Miracle’s SQLServer Open World. An abstract for PASS2008
was rejected last week Probably everybody in the SQLServer world already knows
so much about XEvents and Wait event based profiling..

There is a simple-talk.com paper in the making…