Thoughts on unified integer literal improvements

[Paul Benedict]

Forgive any obtuseness on my part, but I believe the 'y' suffix is
esoteric and unnecessary. I don't remember any other programming
language that has that literal suffix. Is it perhaps something novel?

byte b1 = 0xFF;
byte b2 = 3;

Can someone explain to me why this shouldn't straight compile? Or why
widening must occur to the literal since the type is clearly known to
the compiler? It should be an error if the token cannot fit into a
byte memory space (e.g., 0x100).

Paul