Well.......! I've beating people over the head around here
for more than a year now with the fact that Native SATA will be (is) a better all-around performer than "conventional" Parallel ATA.
Still, there has been many doubters, as they continued to focus on the wrong points -- mainly the disparity between media read/write rates and channel bandwidth. Native SATA is more efficient at the protocol level and at the wire level than Parallel ATA. In the end, what you'll have with a mature SATA is smooth, balanced, "SCSI-like" performance.
As for some of the early SATA performance hiccups we've seen, these are the likely culprits:
1.) Non-native SATA implementations in the form of performance-robbing serial-to-parallel bridge circuits (which also drive up hard drive cost significantly), and/or...
2.) Non-optimal drive firmware -- probably patched up Parallel ATA firmware -- that is not optimised for the operating characteristics of the SATA 150 Mb/s channel.
It also needs to be hammered home that going to Native SATA will actually reduce overall storage costs. Even though a SATA data cable costs a few coins more than a blue-ended ATA-100 cable (now), the Native SATA drive controller circuitry is less complex than Parallel ATA, making it less expensive to manufacture once the assembly lines ramp up production volumes of SATA hard drive controller printed circuit boards.