The opposing viewpoint is "Since we have to use closed source, it would be impossible to do this." This is incredibly myopic. But we can get these programs written faster with closed alternatives.... More hogwash.
The real truth of it is this. If we wait ten more years to pass something like this, there will be an ever increasing number of proprietary apps wormed through our research code. There is nothing so pressing in science that it cannot afford to be done right. (In fact, we should probably slow the march of progress long enough to perfect what we have anyway...)
The author of the opposing viewpoint carps about the millions of lines of proprietary code that would have to be re-written first. Would this be easier in ten years?? No.
The author comments on some proprietary projects of his own he has in mind. Fine, don't produce them with public money.
My belief is that public money should benefit the public. If this had been in place ten years ago, the state of computer programming the world over would be that much better. Just like an author reads Melville, Hawthorne, etc. to learn to write well, so could we study this large public codebase as programmers, and see how the pro's do it.
If the academic world in infested with proprietary software, it is time to eradicate the vermin. Besides, some programmer(s) will just end up getting a grant to re-write the proprietary portions. You can't honestly tell me that this would be more drastic than Y2K prep, or a BSA audit. Or are you against programmers working??