Your Tony Table isn't a Tony Table. Remember, the point isn't just to say "Before: syntax is invalid. After: syntax is valid." The point is to actually show what your proposal does, and what new usage it enables.
IIUC, you're proposing basically this:
So you need to show that this is implementable for compiler vendors. How will the compiler distinguish the mangling and typeid of `S<int, / int, int, int>` from the mangling and typeid of `S<int, int, / int, int>`? To do this, you'll need to know something about name-mangling. If you do not do this part, then every compiler vendor will dismiss the paper immediately as it doesn't engage with the actual issue at any level.
The proposal also needs wording. Right now, it doesn't even mention why existing compilers reject that code! Is there currently wording in the Standard that requires compilers to reject it? If so, where is that wording? Are you proposing to simply strike that wording; or modify it somehow, and if so, how? (Striking sounds good to me.)