1. The sensor senses some light. There's your image data. You can't get to it.
2. And then proprietary camera software extracts that from the sensor, and runs it through proprietary camera hardware and firmware to produce a file format. At no point here can you get to your image data.
3. The file format is written to some sort of media, typically. In, typically, a proprietary format.
4. Then you pull that off the media convert that into whatever you like on an external computer, which you are free to do. Now, and only now, can you get at your image data.
The "cameras should support DNG" essentially moves that last step back by one step, from 4 to 3.
This is some sort of important philosophical distinction? Moving the point at which it leaves the cold hard hands of the camera maker back by one step? If it's truly about having access to your precious bits, surely you should be demanding access to the sensor? To the in-camera pipeline? Why is it suddenly a gigantic issue at step 4, but not at step 2?
This is a convenience feature, dressing it up as a philosophical issue is a red herring. A red herring manifestly designed to distract from the issue of 'well, there are legitimate concerns on the camera maker's side, in addition to the legitimate concerns on the camera user's side'.
If there are legitimate concerns on both sides, and it's a relatively minor feature, then there's room for an actual discussion, and ideas like ROI matter. By making it "philosophical" you take that off the table. With is excellent rhetorical technique, until you're busted. Then it kind of falls apart.