RME is STILL selling that 9632 card - so this is not EOL cycle. This is artificially EOL for Apple users, only. As you say: they actively had to remove code to drop support and that really looks like a b/s ripoff to force us to buy newer hardware.
RME could simply release the source code of that 1.7x driver
but hell I guess they are not going to. And hell yes: I am not going to buy RME anymore. Once I am pissed of a company there is no 2nd chance.
If you decide to temper with the current driver, just let us know, I suppose there are some ppl (like me) willing to test.
GPL'ing the 1.73 driver itself and EOL'ing support would be the honourable thing to do - help legacy users and enthusiasts get every last mile out of their trusted hardware and get an unbelievable amount of goodwill. I don't think there is so much secret-sauce über-valuable intellectual property contained in a coreaudio driver, but getting any audio company to open-source anything is like getting blood out of a stone - it's just in their corporate nature to maintain control over everything.
That said, there is a working open-source ALSA driver ( and even a totalmix clone! ) for linux, so it's not beyond the realms of possibility that some guru with enough driver experience and a bit of time could come up with an open-source coreaudio driver. The internal structure and logic of the xilinx chip must be exposed in that ALSA driver. It's been done before for other hackintosh hardware, why not for RME?