Answer from cs61c-ew (Joo-Rak Son 16103505) for Question 2
I am guessing that if this were the case, there could be two consequences. (1) It may be harder to catch some of the lower-level glitches, such as memory and register allocation. Since the C programmer would only be allowed to access the C program, he cannot have a "microscopic view" at his program and will not be able to find some bugs that are lurking at the very low level. (2) This would be a waste of computing resouces including time, because without the linker, every time the programmer changes a little line of his code, he would have to compile and assemble the whole program again (instead of just doing that procedure independently; pg 159).