Battle Against Any Guess

BAAG - what is it all about? The main idea is to eliminate guesswork from our decision making process — once and for all. It’s not Oracle specific even though the roots lie in the area of Oracle database administration. So keep reading even if you don’t know anything about Oracle or computers. Before we go any further I should say that the title was inspired by another fine BAA… site — BAARF. Just to avoid any guesses.

After some mental fermentation the birth of BAAG had been finally triggered by an Oracle-L thread started as 10gR2 performance sux. Title of the post was a direct consequence of pure guess and, worse yet, uneducated guess. It’s also interesting to see how the thread developed - guesswork. Another problem in the same thread was followed up by the same troubleshooting style. Perhaps, the author, inspired by previous responses, also hoped to find the magic solution from of Oracle-L powered (and often educated) guesses.

Database performance tuning is one of the most noticeable areas where guesswork is still flourishing. There were quite a few methodological techniques born recently, such as YAPP and Method R, and older guess-based methods such as Tuning by BCHR are going away. Nevertheless, people’s mentality is very difficult to change and we keep relying on guesswork hoping to solve the issue faster. Hope-powered guess is the evil.

Performance tuning is just one example. Imagine that your Oracle database crushes with ORA-600 and guesswork starts powered my your experience. Depending how good your memory and experience are you might want to ask in an Oracle forum something like “My database x.x.x.x crashed with ORA-600. Anyone seen it?” It’s fishing for a guess and be sure there will be quite a few people suggesting the fixes right away. If people were not lazy, you would have hundreds of suggestions so consider yourself lucky.

Interesting that one of the biggest problem is — it’s contagious. People don’t even notice as they get involved into guessing competition trying to shortcut a more scientific and reliable investigation path. It’s true that some of the guesses might be correct but implementing a guess-based solution and finding out that it’s useless is quite painful. It undermines the authority of a guesswork follower.

Moreover, often guesswork can be dangerous. Let’s step away from IT and look at, for example, medicine. Imagine how thrilling would be if doctors make a diagnosis based on a guess! Let’s say you are in the hospital after a skiing accident and doctor guesses that your leg is broken instead of doing a proper X-Ray. When I was just 1+ year old, doctors stated that I had Oesophageal cancer and I was about to be operated. Fortunately, my parents insisted on more detailed diagnosis and it came out as a simple non-malignant growth. Imagine you are riding a bike with 100 km/h and in front of you is a blind summit — would you guess that there is no curve right after it?

Do you want to give any chance to a guess in your life? You might discipline yourself but often outcome depends on other people so let’s join out forces to fight guess-based decisions and advocate engineering approaches. Guesswork - just say no!

DISCLAIMER: There is one guess allowed — 42.