A MIMO-enabled endpoint uses two or more (usually three to five) antennas to talk to a MIMO-enabled access point over several channels simultaneously, taking advantage of multiple paths between the devices to augment bandwidth and get around obstacles.

The trouble is that 802.11n is incompletely defined thus far, so vendors must make things up to fill in the blanks spaces in 802.11n's specs. The obvious caveats for any pre-standard product applies: the product likely will not interoperate with other vendors, may not perform as well as the standard anticipates, and may never end up being compliant with the actual standard when it finally arrives.

To take advantage of MIMO's performance gains, all devices in a network must be MIMO enabled. In my tests, any non-MIMO devices dragged the network back down to 802.11g speeds instantly. Given the two- to three-times higher price for MIMO-enabled gear, you have to ask yourself if the cost is worth the risk. There are plenty of people who bought pre-802.11g gear who discovered they had incompatible doorstops when 802.11g became official.

If you're building a new network and can afford to throw the first iteration away, then MIMO might be a reasonable early leap. But for most of us, the fact that 802.11g works very well, and that we must coexist with legacy 802.11b and g devices, makes MIMO a no no.