Considering the number of complaints about hot MacBooks and Imacs, I wonder if they would have trouble with overheating if you started to throw workstation class loads at them.
A TBolt drive is just a SATA drive - drive X in a TBolt expansion won't be faster or slower than the same drive X internally mounted.
For most things an SSD, even hampered by 1394b, would be a lot "snappier" than a spinning drive. For large transfers though, the internal SATA drive would go faster than the peak 80 MBps of 1394b. On any test that would involve head movements, though, the SSD would probably win.
For example, copying a large file from one internal SATA drive to a second internal SATA drive should be faster than copying from the internal SATA drive to the SSD.
Copying the file onto the same internal drive would cause lots of head movements, and the SSD would probably be faster at copying onto the same drive.
and intel cannot include eSata, much cheaper and I believe you can boot from it as well. The rotating HDD is the bottleneck and considering SSD can run faster than Sata III/ 6, providing an eSata port rather than a TB port would make multiple products already available more viable then to introduce a new product I/O to the mix. Can TB run power and handle enough power to run a display or even a multiple bay storage device? If not what is the point, considering that Light Peak is a better way to go down the road when fibre is cheaper. I see another DVD Ram fiasco. 
