Instead of uniformly sampling the key from $\{0,1\}\backslash$lambda $,$ it is more

secure to sample it from $\{0,1\}\backslash$lambda $\backslash \{0\backslash$lambda $\}($i$.$e$.,$ exclude the all$-$zeros key$)$ as it

prevents the plaintext from being sent in the clear when using OTP