Errata in Homework 1, Due Feb 7

The versions of the homework now posted on the web page include these corrections.

In Question 1, part 6, the question is only when xlower and xupper are adjacent IEEE floating point numbers, not other floating point numbers. In particular, they are binary.

In Question 4, part 3, the last sentence should have asked you to confirm numerically the same error bound that the first sentence asked you to prove. In particular, a factor of exp(2x) was missing from the error bound in the last sentence.

This is not a correction but a clarification: in Question 4, Parts 2 and 3, if you find it convenient to prove slightly smaller (better) upper bounds than the ones given by replacing the factors 3i-2 and 3(i-1) by something a little smaller, go ahead.