The Law of Logical Argument…

I received a mail today containing a list of laws (in the scientific sense), all of which were silly observations just for amusement, but this one stood out.

“The Law of Logical Argument – Anything is possible if you don’t know what you are talking about.”

Although not directly relevant, that got me thinking… I answer a lot of questions on forums and one of the things that strikes me time and again is people find it really difficult to define and communicate their problem. In some cases there’s a language barrier to contend with, but most of the time it comes down to a basic lack of thought by the poster. How many times have you spent days following a forum thread only to find the question you were answering was not the question the original poster “thought” they had asked? This is not just a feature of forum posts, but also in every day interactions. People will spend hours skating around issues, but never formally outline what the issue is. In many cases, the process of clearly defining the issue actually allows them to answer the question for themselves, or at least highlights a possible path of investigation.

Unless you can clearly define a problem, there is no way you can communicate it to anyone else. What’s more, unless you have an understanding of your problem, you can’t judge the value of answers, making you susceptible to suggested voodoo solutions.

Over the years I’ve had the good fortune of interacting with a lot of people I consider Uber Geeks, but what continually strikes me is the simplicity of their approach to solving problems. They may sometimes look like they are making leaps, but that’s just because experience allows them to do the steps quickly in their heads, rather than have to draft it out on paper, but most of the time they are following a fairly simple recipe to solve the problem, the first stage of which is actually defining the problem itself.

Is it surprising that database performance tuning is still considered by many to be some form of black magic? Not at all, because it is one of those areas that requires thought, reason and a meticulous approach.

Enough ranting for today. I’m off to run a database on a single 1TB disk and wonder why my I/O throughput is so bad. Must need a faster CPU. :)

Cheers

Tim…

PS. Someone on Twitter recently linked to an article called The Lost Tools of Learning. It’s quite long an a bit heavy at times, but some of the points it makes tie in with this post and I think it’s worth a read.