You make some good points. The only reason that I linked to a sustained transfer rate test in the first link is that If the drive cannot do a sustained transfer faster than what the optical drive needs, then the burn is going to have problems. It does have to be able to do a sustained transfer fast enough, but like you said, there is more to it than that. If it is a boot drive (which I kind of assumed it wouldn't be as he said it was going to be a second drive), you have to be able to randomly read as the os and other programs (multiple background program's, even if you don't have programs running) will be accessing the derive at the same time, including the burning program. Even still, a 7200 rpm drive as boot drive should be able to do a single 16x burn without problems. There is even less of an issue if it is a data drive.
as far as a raptor being better and faster, to an extent I agree that access times and such are very important, particularly with a boot drive. A raptor tests with a sequential read rate of 69MBps. The 3 drive array that I run has a sequential read rate of 144MBps (207Mbps buffered). You would think that would mean that the 3 drive array is way faster (and in certain circumstances it is). If you are running a boot drive though, that fast access time of the raptor kills the difference. The raptor has a random read rate of 49Mbps. The 3 drive array that I am running has a random read rate of 63Mbps. That huge speed difference is gone due to the fast access time of the raptor (though the raid array does seem to compete and is cheaper per gig). It wouldn't surprise me if a 2 drive raid might be the same or a bit slower than a raptor (though getting close to raptor random access speeds for a lot less per gig might appeal to some).
All in all, I only look at these benchmarks to get a rough idea anyway. You can run a benchmark 5 times and get different results each time (though they tend to be in the same ballpark). The controller can make a huge difference too (which would apply to raptors too). I have a true 4 channel pata promise raid card (not a cheap card). In a 2 drive raid, it can barely beat a single drive. In a 4 drive raid, it barely beats a 2 drive raid on my much cheaper siig raid card (which rather pissed me off). I'm actually quite amazed by the three drive raid that I am running now (I have only been running it for a short while and it is by far the fastest I have seen). The weird thing is that a 4 drive raid on the same controllers is slower (I'm guessing due to the drives). I have 3 matched hitachi sata drives, and a single pata hitachi. I tried it with a syba pata sata adapter and the forth drive slowed it down. I may have to test it with my abit sata pata adapter as it ran great with raids in the past (but wont fit in this case, dang it). I have even noticed that single drives can have speed differences on different controllers. All in all, there are endless variables. I really don't want to argue weather a raptor or a raid array is faster, as there are so many variables. I do think it is fair to say that taking cost per gig and speed into account, raid arrays can compete in a certain respect with raptors, no problem. Raptors surely have their advantages too, though they are a little pricey for me (especially considering the performance I have gotten from some raid arrays). If nothing else, a raid 0 with 2 7200 rpm drives will beat a single 7200 rpm drive hands down, and cost per gig is similar.