I have designed Randy Rom in 1996 for MLC in Moers, Germany. There are four jumpers; the group of three is to cut the reset line from the three IDE ports - this was necessary with kick 3.0 and some harddrives. This is not a very clean solution, but it worked at the time. Kick 3.1 solved the problem on the software side. If you have kick 3.1, close all of these three jumpers.
The one jumper that sits alone there selects one of the two "1" ports, I think it's labelled "on=b". The "a" port is the 2.5 inch port, and the "b" port is the 3.5inch IDE port. The outer port is always active. This jumper was necessary because the 2.5inch port is not buffered.
The interface is fully IDE-fix compatible. It was licensed to MLC. That company has sold it with illegal copies of IDE-fix. When they were found guilty in court, I have cancelled the contract with them, but they continued making and selling the interface - that's why I consider them illegal.
Don't worry, I won't go after anyone using this adapter. I have found and confiscated the overstock of these adapters a few years ago, so they're out of the market. I might even start selling this overstock in the future, because there's currently no IDE-fix compatible buffered adapter available from resellers.
If it makes you feel better, PayPal me a few bucks ;-)
Your third question can't be answered that easy, because I have different reports from different people. Some claim that a vanilla A1200 boots fine with such an interface, others (including me) have different experience. It seems like it all depends on the CF card itself. These cards have a so-called "true IDE mode" in which they emulate IDE harddrives. Depending on the accuracy of that emulation, the A1200 will boot, or it won't. The adapters are all the same, and should all follow the schematics that Sandisk has published, so you can choose the cheapest for your experiments, knowing that a more expensive one is exactly the same.
Jens Schoenfeld