Sorry about the "about what" comment James, I was being a smart-ass at the time.
The MIDI spec spells out all sorts of things, one of which is that a standard MIDI device must be able to communicate over a MIDI port at 31.5 kBaud. If your Korg can do it faster over a "to host" port directly to a computer, it doesn't break my theory for two reasons - one, it's not my theory. It's the MIDI 1.0 specification. and two, that's technically NOT a MIDI port, it's a direct interface using MIDI protocols. A technicality, to be sure, but a defensible one. As near as I can recall, the MIDI 1.0 spec states that a device must be able to sustain 31.5 kBaud. I don't remember it stating that devices couldn't go any FASTER than that, and you're right about my not mentioning that earlier.
The thing is, any time you start mixing and matching devices within a certain specification, you have to assume that at least one of those devices will BARELY meet spec, or perhaps even FAIL to meet the spec. Therefore, in order to predict what kind of performance the system as a whole will have, you have to be pessimistic or expect to be dis-appointed.
As far as the link you posted, these devices can come in handy if you have the particular problem they solve. The thing to remember is that any electronic device has a finite amount of propagation delay by it's very design, and if you put too many of them in series those delays can get large enough to be noticeable. Still, if all you need is for one output to go to three different places, or for three different MIDI streams to go to the same input, then a simple box can be the best solution.
Midi solutions has been around for quite some time, and I don't recall ever hearing any negative feedback about them. I haven't personally used any of their gear, but wouldn't hesitate to do so if I needed a particular function that they could do... Steve