I think that article is pretty crazy apologetics for a bad mechanism. There's an elegant mechanism design solution to the inefficiency inherent in HFT:

http://blogs.reuters.com/felix-salmon/2009/07/31/solving-the...

See also the first paragraph of this: http://messymatters.com/2009/07/31/wellmanblog/