Today is Saturday, 16th of June 2007 and is the birthday of… (drum-roll please…) …BAAG - Battle Against Any Guess.
This idea was sitting in my head for a while and finally materialized here. “Enough is enough” trigger was this Oracle-L thread followed by couple emails from Tanel Poder and David Kurtz. While replying to Tanel describing this idea, I carried on and imperceptibly for myself started this site.
Interested? Find more details here.
Find more about baag
Not guessing at all would imply there is always a fool-proof method to determine exactly what the problem is. But there’s no such thing thing, is there? You may for example, determine that a dictionary query is slow due to a FTS instead of doing index-lookups. But you wouldn’t know exactly why it’s doing that until you guess at a few possible causes, and try out a few solutions. Especially if it’s caused by a bug. It’s no different when searching Metalink for possible solutions, and guessing what you have found (and it’s recommended solution) may be the answer to your problem. Sometimes Metalink may provide a method for determining if it is the same problem, but many other times you just have to guess. I see where you are coming from, and while one should avoid wild guesses, making a educated guess is a valid method in problem solving.
Thanks for your comment Ben.
In this case, there are options to investigate that. Almost certainly it would be CBO estimating cost incorrectly or Oracle performing access to a fixed table ineffectively. Different techniques can be applied here — 10053 trace, hinting and comparing the costs, 10046 trace, etc.
Searching Metalink and Google is a great idea but it might result in completely irrelevant fixes because people are looking for a ready to apply solution as if Google and Metalink can diagnose the issue instead of an analyst. Google/Metalink should rather provide more details to aid the research.
Educated Guess? Often it’s not easy to distinguish educated guess from gut feeling or simply wild guess. Assumption is a good term and it must be verified before relying on it.
“Imagine you are riding a bike with 100 km/h and in front of you is a blind summit”
Actually, a condition known as “Death wobble” can kick in at around 80 kmph. I experienced moderate wobble at 50.9 mph descending PA St Rt 31 just east of the Laurel Highlands. So I can’t imagine riding at 100 kmph under any conditions on a bicycle, thanks.
CrazyGuyOnABike
You must be right and I should have said 42 mph.
Anyway, welcome to the club!
Alex,
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.