When new software comes out, it wasnt just slapped together... theres bound to be bugs left and right in the source code. The only way to make sure that your product does not have any glitches or bugs is to have people test the product before it hits the market. These people are given the task of running it through its paces, almost striving to find any problems. The main key is that a piece of software may run great on one machine, but another machine with a different hardware configuration may come across a problem previously unseen. This is then reported to the programmers, and they set out to remedy the problem. As for the need for a large amount of people, well, it simply increases the chance of finding problems throughout computers with different OS's, processors, vid cards, etc. And along with some blood, sweat, and tears, that is how you design worthwhile software.