First question to ask is what is your budget and what is your need?
If you have strictly utilitarian needs, and gaming is a secondary/tertiary need, shop for a $300 to $400 laptop computer first, and see if it does what you want. Even if you want a desktop, keep that figure in the back of your head that you can get a reasonably featured Intel i-series laptop for around $400. That's a good guide to determine if your build is reasonably priced.
Second, look around the 'net and magazines for recent custom build that people have done. People love to brag about their builds. Try to talk them up about any issues they ran into with their specific build as that will tell you what components are really better, and which components might have conflicts with each other. I know it seems lame compared to going to school or buying books, but it is market research that actually pays off because it saves money. It's the kind of info you won't find in a book, because it's a little too true to print.
At any rate, price some of the builds from a reasonable supplier like Newegg.com or MicroCenter, or whatever has a good rep in your region. Decide what you can really afford versus what you want. Realize some things are overpriced per performance gain, and there are bottlenecks in computer design such that spending tons of money on one part while being cheap on a component that interacts with the other part can actually just be a waste of money in sum. Don't be fooled into thinking the newest or most powerful component is actually the better value. Also, figure that you should get paid about $30/ hour for your installation work and factored that into your final cost.
THEN, do the reading (books or online) on:
1. cases, motherboards, cpus, RAM, videocards, etc... Youtube actually has some decent videos on how installations work.
2. case cooling and power supply issues. Running at reasonable temps is the best thing for getting the most out of a build.
sidenote: Don't pay attention to overclocking and overclocking needs. Don't bother doing it. If you want to learn about overclocking, practice with an old computer after you build a new. Overclocking CAN kill a computer by over-volting and overheating parts. Reading about overclocking from someone who know the physics and electronics of it in detail is informative, but it's risking your money to actually overclock.
3. diagnosing computer problems. This actually is pretty helpful to have done before you do your first build, as bad components, or bad practice might happen. I've had several good builds, and then a few bad ones where I did something wrong that I forgot about though I knew better earlier. Being good at troubleshooting saves a lot of stress when you build by yourself.
A book that teaches A+ cert concepts might be overkill, but there's a Dummies guide version for around $50 ($5 used on Amazon), and that can actually be helpful for background info, although my first build was done with nothing more than tips from a Maximum PC special issue. You don't need much more than that then some physics 1/2 appreciation for electricity, circuits and a run-down on how computer hardware is and works.
At any rate, don't actually build or buy components until you feel you've learned enough that you could do it from memory. That'll prevent you from doing something stupid and wasting money. Or if you feel playful, build a $200 computer from the cheapest/used parts available to get practice in. Then overclock it to death.
My current build (and temperatures, not under load):
Operating System
MS Windows 7 64-bit SP1
CPU
Intel Core i5 2500 @ 3.30GHz 44 °C
Sandy Bridge 32nm Technology
RAM
16.0 GB Dual-Channel DDR3 @ 665MHz (9-9-9-24)
Motherboard
Intel Corporation DZ68BC (CPU 1) 39 °C
Graphics
SyncMaster (1280x960@60Hz)
1024MB GeForce GTS 450 (EVGA) 43 °C
Hard Drives
244GB Seagate ST3250310NS (SATA) 38 °C
977GB Western Digital WDC WD10EACS-00ZJB0 (SATA) 35 °C
Optical Drives
ASUS DRW-24B1ST
Audio
Realtek High Definition Audio
