I know this topic has been addressed in detail elsewhere (see goofy’s post here, for example), but a friend recently asked me about it, so I thought I’d take a crack at it. It’s fairly straightforward: there are the complex, implicit rules that people have been following for over a thousand years, and then there are the simple, explicit, artificial rules that some people have been trying to inflict on everyone else for the last couple of centuries.
The explicit rule is this: use fewer for count nouns (things that can be numbered), and use less for mass nouns (things that are typically measured). So you’d say fewer eggs but less milk, fewer books but less information. Units of time, money, distance, and so on are usually treated as mass nouns (so you’d say less than ten years old, not fewer than ten years old. One handy (but overly simplistic) way to tell mass nouns and count nouns apart (save for the exception I just noted) is this: if you can make it plural and use a numeral in front of it (five eggs), then it’s a count noun and it takes fewer.
The only problem with this rule is that it was invented by Robert Baker in 1770, and it contradicts historical and present-day usage. In actual practice, fewer has always been restricted to count nouns, but less is often used with count nouns, too, especially in certain constructions like twenty-five words or less, no less than one hundred people, and one less problem to worry about. It used to be that people used less when it sounded natural and nobody worried about it, but then some guy in the eighteenth century got the bright idea that we should always use one word for count nouns and one word for mass nouns, and people have been freaking out about it ever since.
Baker’s rule is appealing because it’s simple and (in my opinion) because it allows people to judge others who don’t know grammar. It makes a certain kind of sense to use one word for one thing and another word for another thing, but the fact is that language is seldom so neat and tidy. Real language is full of complexities and exceptions to rules, and the amazing thing is that we learn all of these rules naturally just by listening to and talking with other people. Breaking Baker’s rule is not a sign of lazy thinking or sloppy grammar or anything else negative—it’s just a sign that you’re a native speaker.
The fact that not everybody follows the simple, explicit rule, nearly 240 years after it was created, shows you just how hard it is to get people to change their linguistic habits. Is there any advantage to following the made-up rule? Probably not, aside from avoiding stigma from people who like to look down their noses at those who they deem to have poor grammar. So if you want to please the fussy grammarian types, be sure to use follow Baker’s made-up rule. If you don’t care about those types, use whatever comes naturally to you.