A word of warning for your consideration: the reference given by Diodorus Siculus is very good - clear, detailed and covering most aspects of the construction and application of PC RAID systems. In my opinion, however, it tends to overstate the advantage of improved speed and neglect the disadvantages of RAID0 arrays.
In real-world tests (as opposed to theoretical calculations), the speed advantage for most applications is barely detectable, including games for example, where you might expect to see a significant improvement. This is generally due to the fact that other factors play a significant role. There are indeed applications where a gain of up to around 40%, depending on the RAID controller used, can be achieved but these are relatively few and far between.
The main disadvantage of RAID0 is that it is not a genuine RAID system - no redundancy - and the data is striped across both disks. So if a fault affects one of the HDs, you lose all of the stored contents. Moreover, it is very difficult to recover data from the failed array - a job for the professionals only, at considerable expense. So it is essential that every file of any importance is backed up regularly and reliably (not that you shouldn't do this for any storage system, but many users seem to rely on data recovery techniques which at least stand a chance of working for normal and RAID1 HDs).