Well that's a good writeup, except that the person who challenged you did so for the wrong reasons.
The IDE config "debate" is not about a single device becoming slower if it is in a master/slave configuration. It has to do with copying, ripping, or burning to devices that may share the same channel, as on the IDE bus only one device can talk at once. I see that at the end there you do cover that point. This is also why disabling auto insert notification on the drive sharing with your audio drive is recommended, although the actual problem won't manifest itself until you're taxing the system.
Another thing you might try is putting a CDROM drive in PIO mode and then seeing if that doesn't slow down a hard drive on the same channel. I believe that was the original IDE issue from long ago you spoke of. Also, after much testing of my own long ago, I concluded that sustained transfer did not increase as max burst speed increased...thus sustained transfer doesn't improve as you move from ATA/33, 66, 100, etc. Those tests were run long ago, however, so I can't say for sure whether that is still true. My point being that you can't tell what mode a device is operating in simply by looking at sustained transfer rates.
Good writeup!
Slackmaster 2000