In the world of music technology, par for the course would be for protocols like this to be purposefully obfuscated, and constantly reinvented. Good on Korg for seeing the value of releasing this.
This is cool and reminds me that when I bother to think about it I'm continually surprised no major camera maker has released a camera with an open source firmware.
Granted, I don't think end-users care about the openness of their cameras, but it only takes a small core of developers who do care to donate a lot of time and vastly increase the value of your device. As an example, if you hang out on forums for Canon cameras you will find a lot of people who factor whether Magic Lantern (and open source hack on top of closed source Canon firmware) will run on the new model or not.
I particularly wish Sony (who is killing it with their sensor designs lately) would jump on this.
Korg have literally always been amazing about stuff like this. They were the first in the App Store on iOS, with an iPad app on day 1 for the iPad's release. Their MIDI implementations are always more consistent, more solid, and more reliable than competitors.
If you're surprised by this, you might know music manufacturers in general, but you don't know Korg. :-)
This kind of thing in particular seems to be a first for them, seeing as there are no other 'korginc' repos on github. I do hope they publish more there.
I'm not sure what exactly you are trying to say about their MIDI implementation, but fwiw I have had less headaches while implementing things based on Yamaha's SysEx documentation than Korg's. Not that either company seems to be trying to hide things at the MIDI/SysEx level.
Not inconceivable the two are connected somehow, of course (either in tribute from Korg or even the other way round; AFX has released a bunch of tracks named after gear, e.g. Fenixfunk 5... and it'd hardly be surprising if certain artists had access to pre-launch gear.)
