Yes its standard practice, but no it doesn't produce the best strategy in the context of 802.11. The problem is that the 802.11 MAC clear channel assessment minimizes the opportunity for interference but in fact reduces the opportunity to transmit even in cases that won't interfere. That is I might be trying to communicate with an AP that is 10 metres away, but can see a transmission on the same channel from an AP 100 metres away and the CCA will prevent me from transmitting, even though I probably won't interfere with the other transmission and he certainly won't interfere with me.
The CCA in 802.11ag is defined in two different manners. The first is that if I can synchronize with a preamble in the channel then I consider the channel occupied. This basically means that anything I can hear at the receiver sensitivity (-85dBm for 802.11ag) in my channel I won't transmit. However this doesn't help for the issue of overlapping channels or non 802.11 transmissions in band. So the second definition is a simple RMS power measurement at 20dB above the receiver sensitivity level. That is if I hear something at -65dBm in my band I don't transmit regardless of what type of signal it is. So I'd rather have someone transmitting on the same channel as I am as far away as possible, and adding more, but overlapping channels, giving the frequency planner more opportunities to do that.
Therefore a better frequency plan in the 2.4GHz band that gains a 20dB advantage from the CCA definition is 1, 7, 13, 2, 8, 14, as that minimizes channel overlaps while maximizing the distance. However channel 14 isn't always available and so 1, 6, 12, 2, 7, 13 is a compromise that is reasonable.
Note that the 802.11n "green-fields" preambles have a better definition of the CCA that is basically the same whether or not your on the same channel. However, I don't think anyones really using these new preambles in 802.11n yet.
D.