Defrag (pt 1) You can't measure it.
I can't understand the fascination of most otherwise perfectly normal Windows users with the Defrag option in Windows. It's reviled as not being powerful enough, it's credited with all kinds of performance improvements and even suspected of concealing superpowers to fix all kinds of problems.
I don't understand why. I can't think of any other community of users who are so obsessed with how 'fragmented' their computer hard disks are. And the problem only seems to be coming to a boil once more with Windows Vista, where Microsoft have taken a few hard decisions in their redesign and update of this operating system background task.
No matter, like all good bloggers, I never saw a bandwaggon I couldn't jump on, so I'm actually going to make an effort to address some of the more common myths and questions around this part of the operating system.
Part 1 - how do you measure 'fragmentation'?
Obviously by asking your defragger of choice if it considers your hard disk to be fragmented or not. But what if you don't trust your defragger for some reason? Now most people would go and get a life, but for those of you who are in the grip of Windows disk fragmentation paranoia, this isn't enough. What you must do instead is purchase several commercial defraggers and use them to test each other.
Then you get very unhappy because you can't seem to get your disk defragged properly. You run one product and let it do what it wants and when it's finished you test it with another product, and that finds some fragments, and so you run this second program and test it with the first one and go around in circles until you give up and ask for help. Why?
All the different products have slightly different ideas about how to measure fragmentation, what the best way is to defrag a drive, what the optimal settings are to mitigate against easy fragmentation in the future, etc. Hence, what one product calls "defragmented" might not appear so when the disk is examined with another product.
This doesn't mean that one product is right and the other wrong. It's more like... well lets say you measure something with a ruler and tell me the measurement is "20" but not whether you're talking about inches or centimeters, or something else entirely.
Are you "wrong"? No. But can I actually use that number for anything useful? Again, No.
In other words, you can't reliably measure and compare fragmentation between two different products. They've both got different ideas of what exactly the problem is; different ideas of what to measure to diagnose the problem, and how to do it; what exactly should be done to cure it; and sometimes those ideas are in direct contradiction to each other.
The more cynical people in the audience (or those who use other operating systems and haven't contracted this paranoia) may be asking themselves just how much of a problem a fragmented hard disk is, and if maybe these products are a prime example of snake oil.
Frankly, I'm not sure that the products are snake oil, but some of the hysteria surrounding the way these utilities are sold to 'home user' types does make me wonder. In some special cases, workstation drive fragmentation can be a real issue that needs to be carefully addressed, and servers should have all aspects of drive health checked on a regular basis, but for most home users the built in tools provided by Windows will be more than enough to keep things running well.