I have worked as an IT consultant for the last several years, more recently with my own company. Some of the tasks handed to me during that time seemed at first to be impossible to solve. There was just too much to grasp. It occurred to me that this was no different than other types of work I had done, but was merely a similar machine composed of different parts. My work for the 10 years previous had been in electronic repair up to including the repair of industrial and broadcast television equipment in the days when computers were beginning to rule there as well. A fair amount of the work done in those areas had to do more with mechanical issues rather than electronics.

 Sir Isaac Newton is widely recognized as the father of the scientific method. In short, it is a collection of methods used for the organized discovery of and cataloging of scientific phenomena. I have to admit that what little I do know of it was learned in elementary, middle and high school science classes, and more recently from Wikipedia. However, in consideration of the time having passed before my education rather than since and the fact that what I learned was essentially from Newton and his ilk, what I have managed to grasp and maintain should be close enough for the purpose of this article.

 Use of certain methods has helped me greatly in various fields and endeavors. The concepts are not particularly difficult. But it seems as though they are often overlooked by many. Troubleshooting any problem requires a certain amount of discipline and, given time to work on a problem, a thorough examination will be the most effective in determining the proper solution.

 So, what was I to do now faced with a broken Exchange 5.0 server? No wheels, pulleys, belts, or transistors, and certainly no bells or whistles. My previous experience with Windows had been limited to working with Windows 3.0, 3.1 and Windows 95. Fortunately, I had also worked with Linux for several years. The first order of business was to not get lost at the start.

Focus…

 First, only look at smaller pieces of the big picture. Each piece will be used to build an overview of the actual situation or problem. Trying to evaluate the problem by staring at it leads to, well, staring at it. Hard to make progress this way. It is too easy to get bogged down looking at the big picture.

Figure out what works!

 Whatever the problem or symptom, I find it is best to first check the rest of the system and see what is still working. Sometimes, the failure might be due to some other underlying issue, and clearing that up first will magically fix the primary complaint. Just because everything else is so-called working doesn’t mean that a quick restart of the server or workstation will fix it. Sometimes, as you may know, it can make things worse. Things like improper network setup including and especially DNS can cause other services to fail or worse yet become flaky. A corrupted boot sector won’t necessarily crash a machine, but a reboot could expose you to embarrassment, and your customer to downtime. Getting those things working as they should is critical to avoiding problems in the future.

Establish an overview

 Part of what you will get out of seeing what works rather than kicking the problem in the shins, or worse simply staring at it, is the establishment of an overview, not only of the problem but of the general health of whatever system you’re attempting to repair. Also, with a correct overview of the situation, you can consider options for repair based on available time, cost of downtime, and cost of the repair. Now is a good time to get with the technical rep or office manager for your client. Make sure they understand the extent of the problem and how much time you think it will take to fix it. Also, give them options. They will feel a lot better if they feel like they had a hand in understanding and fixing the problem. Don’t give them options you do not feel comfortable with being able to come through with. Whether or not you should select the option most profitable is outside the scope of this particular article. Just remember, first do no harm is a pretty good professional rule to live by even if you’re not a physician.

Now for the attack!

 Based on feedback from the customer, as well as your collected information and overview, it’s now time for using the scientific method in attacking the problem:

The control..

 The control is something that is known that remains unchanged through troubleshooting. The control should be something you leave alone or monitor while making other changes. Don’t get stuck on your selection of a control, but always have at least one. The object of the control may change in the process, but only if your current experiment proves fruitless. If the control changes without your direct modification and while you are making other adjustments, then you have made a discovery. If it remains the same then it can still be yet one control of many. THE control could be A control. For example, the tint control on a bad television set might also change the color saturation, or vice versa. If it does, then you should refocus your experiment to the color circuitry rather than the rest of the set. If not, then the tint knob can be placed at center and be considered a control in scientific terms.

The experiment

 Once you have a selected a control or something to monitor, you can make changes to other parts of the system. Don’t forget your focus and overview, and don’t go changing the control willy-nilly. The troubleshooting process should be a collection of controlled experiments. Without organized experimentation, your attack will become haphazard. Who knows? You might get lucky. But without careful examination, your luck will not teach you anything. You won’t learn how to fix the same problem in the future, or if it mysteriously repeats itself on the same equipment. You will no doubt end up with several experiments and results. Some will have worked as expected, while others may have done nothing. Still others may yield surprising results. All of these results are important.

Discovery

 As you are making changes and monitoring your control, you should begin to notice results of the changes. Either the control itself changed, or something else changed.

Notes

 If you have an excellent memory, then you can make mental notes. But, regardless of your prowess, be careful to take good notes. For most of us, this means actually writing them on paper. Personally, even though a text editor may save time, actually writing the information down can help to memorize them as you go. These notes might actually help you in the future. Whatever your preference, you should save your notes in some form which may be easily retrieved. A knowledge base or even a personal wiki application will prove invaluable in the future. Properly filed paper or text file notes can also help.