What is 'defragging', and why should I do it?

To you and me, a file on your disk is a single 'thing'. You open it, you work on it, you save it; it's a single entity. Like, say, a book. To your computer, however, a file is a lot more like a bunch of pages in that book that it has to keep track of individually.

Let's briefly define a couple of concepts before we go further. Under Windows a hard disk is nothing more than a collection of information buckets called "clusters". Each cluster is a fixed size, typically 512 bytes or characters. When you create a file on disk, Windows assigns enough clusters to the file to hold it. So if your file is one byte long, it gets one 512 byte cluster. If your file is 600 bytes, it gets two - one 512 bytes full, and one with 88 bytes of data, and 424 bytes unused.

Clusters aren't required to be next to each other on the disk. In fact, that's part of what the "Random" in "Random Access Storage" means; data can be accessed and stored on the disk in random places. So when Windows creates a file, it keeps track of which clusters make up the file, wherever on the disk they might be, and in which order they should go. Kind of like numbering the pages in a book.

Now, imagine if you had the pages of a book randomly distributed around your house. You know where they are and in what order to read them, but you have to run all over the house as you get each successive page.

That's a fragmented file. The clusters that make up the file are scattered throughout the disk. The result is that when you access the file, Windows has to race all over the hard disk to retrieve the whole thing. That takes time.

If instead the pages of your book were all next to each other, in order, then they'd be much easier to read. No need to run all over. That's a defragmented file: all the clusters allocated to the file are in order and physically next to each other on the hard disk.

Files become fragmented because of the way clusters are re-used and allocated on a hard disk. If you delete a file that takes up two clusters, and then write a file that takes four then the new file might be split - two clusters where the old file was, and two clusters somewhere else entirely. Multiply that scenario by thousands of file operations and deletions on your disk every day, with much larger files, and you can see that fragmentation can add up very quickly. The result is your machine gradually slowing down.

Defragging your hard disk is easy. Right click on My Computer, select Manage, and click on Disk Defragmenter. Click on the hard disk you want to defrag, and click on the Defragment button. Defragging can take time, but you'll be able to see the progress as the graphical display of your hard disks state is periodically updated.

Rather than doing it manually, though, if you leave your computer on there's an easy way to schedule the defrag to happen in the middle of the night.

Fire up notepad, and enter the following:

defrag c: >c:\defrag.log 2>&1

Now save that as "c:\defragit.cmd".

Now, in Control Panel, Performance and maintenance, Scheduled tasks, create a new scheduled item. It should run c:\defragit.cmd, and I'd suggest doing it every night when you're not using your machine. Check the log every once in a while to make sure that the process is happening as you expect.

You can run defragit.cmd at any time yourself, if you like. Just open up a command prompt and type c:\defragit.cmd.

Video

Here's a video I put together that walks through some of the steps above: