When I started constructing crosswords for the Times, friends began to ask me what constitutes "cheating" when you're solving a crossword puzzle. I'm not sure why making the puzzles (or even being a regular solver of them) renders me an expert, but I've tried to supply answers over the years, and I've gotten to the point where I have a usual response that I'm comfortable with.
First and foremost, I think the answer is up to everyone as an individual. Crosswords (and other puzzles, or at least ones where there isn't money, pride, etc. at stake), are a competition between you and yourself. While it's trite, it's true: If you "cheat," you're only cheating yourself. That said, I've set a couple of guidelines for myself, and those I've told them to tend to agree with them.
I'm totally happy using the people around me as a resource. It makes the puzzles more social, and, while it's a help, it doesn't trivialize the enterprise. Also, there's a separate, and enjoyable, skill set involving knowing who will know what. I know that if there's a clue about the goalie on the 1973 New York Rangers, my father will know the answer, but he won't have a clue what "http" stands for.
I never Google. I feel like googling takes away the fun, and turns a mental exercise into a tedious task. Google will know the answer. Asking Google is basically the same as reading the answers the next day, something I also never do. That said, this wouldn't be a good law-oriented blog if every rule didn't have its exceptions. If I'm stuck on a clue like "'Catulli Carmina' composer" (From the Saturday, August 6, 2005 NY Times puzzle), then I really just don't care what the answer is. If I google it and find out the answer is "ORFF" (Carl Orff, that is), I'm in no way satisfied -- I'm not smarter, and I haven't learned anything interesting. But some clues obviously impart some interesting piece of information. Given the clue "Asian leader who had a Ph.D. from Princeton," (July 29, 2005 NY Times), while I don't know the answer, I know that the answer ("RHEE") will be interesting. Or other times, it's not an interesting fact, but a bit of word play that has me intrigued. If I have the clue "Bread box?" (also August 6) and have filled in the letters A_M, but can't figure out what the middle letter could possibly be, I'll check back the next day, because I know that I should know the answer, and I know I'll find it clever. (The answer, of course, is "ATM.")
I bring this all up because people have been leaving comments and emailing me about when it's acceptable to use Google on one of my VC puzzles. Lately, I've been trying to include a note saying not to use Google when I think it's inappropriate, but the Google/non-Google distinction is finer than that. With some puzzles, all the research in the world won't help you. You just have to rely on your ingenuity. Others, like this one, clearly require research. The point of that puzzle wasn't to test people's mental images of every nook and cranny of the borders between the 50 states. And then there are the puzzles that fall in between. In this one, if you simply googled the list I provided, you'd quickly find the answer. But I didn't expect anyone to know off-hand the order of ticker tape parades in Manhattan. I certainly didn't expect anyone to be able to identify whose parade was missing from the list. The thought process that I had in mind for solving it was something like: "1) These are mostly New York-related; 2) They are all people/groups who did worthy things in the last several years; 3) They're listed in the order in which they did these worthy things; 4) People who do worthy things in New York get parades; 5) Let me find a list of ticker tape parades and see if this matches up; 6) Not only does it match up, but the missing one -- the Discovery astronauts -- would be particularly appropriate today since the Discovery landed."
To put all of that in a nutshell, I think a good rule of thumb is that you should do whatever research doesn't make the puzzle trivial. If it's a "What do these things have in common"-type puzzle, then don't just google the list. If it's "Who won the Best Supporting Actor Oscar in 1975" (which I promise never to ask), don't go to the IMDb's list of Oscar winners. Etc.