It's a concern, brother. A fundamental issue in all these "compatibility software" projects is: "Why bother shoe-horning proprietary software into Linux when you could apply your resources to further developing our existing open-source tools?".
Yes, you are completely correct on this. I would, personally, never bother to spend a lot of time trying to be compatible with something proprietary.
Another good example is a utility I wrote for mma, ys2mma, which attempts to convert Yamaha style files to something that mma can use. It works, mostly.
But, now Yamaha has a new format for its newer synths ... and ys2mma has real problems with these files. I guess one could spend a lot of time reverse engineering, etc. But, it's really not worthwhile. The formats/concepts really aren't the same to start with, so any conversion will be, at best, "okay".
And, the real problem with this, as you point out, is that we'd be much better off spending our time writing good library files for mma. When I wrote ys2mma I thought it would be a tool which would extract data from an existing file, convert it and then someone would use that to write a proper/good mma file. Unfortunately, the truth turns out to be that once a simple conversion is done ... and that's the end. The results, while usable, do nothing to show off the original file or the conversion. Kara might want to jump in here ... he's done a lot of conversions, but from some recent mails from him (and contributions!) I think he's come to the same thought as me: It is easier to do it from scratch.
The BiaB converter is similar. It extracts some of the info from a file and presents it in a mma format. If you take this as a starting point and develop a good mma song file, that would be wonderful. But, I suspect most folks will just use the translation and be happy.
Oh well, that's life