Don't forget to protect the aldehyde (by forming a Schiff base adduct most likely) if you choose to use n-BuLi route in the above reaction sequence. Also, I ran across this reaction, which may turn out to be quite useful, in a new book called "Compendium of Organic Synthetic Methods Volume 10" (2002) by Michael B. Smith today. It gave the following example reaction and reference:
benzene --> phenol (using 1. e-, TFA/CH2Cl2, 2. triethylamine in 72% yield)
Fujimoto, K.; Toduda, Y.; Maekawa, H.; Matsubara, Y.; Mizuno, T.; _Nishiguchi. I._; "Tetrahedron Lett.," 1996, 37, 3623.
As I have not looked up the reference article yet, I don't know whether this reaction can be successfully applied to vanillin yet or not. However, if it can, then that would be a significant improvement over using n-BuLi to hydroxylate the #5 position, I would think.