This brings up an interesting question: Just what do we mean by "crap"? Are we talking about badly-designed pages, or about search-engine spam?
A photographer's page of up-close-and-personal Emperor Penguin photos might be badly designed, with uncompressed 300Kb JPEG photos, but to the person who's really interested in Emperor Penguins, waiting for a dozen photos to download might be a small price to pay for seeing first-rate, detailed photos of Emperor Penguins in their natural habitat. Certainly the page wouldn't be "crap" in the eyes of an Emperor Penguin fancier, though it might be annoying to penguin fanciers who have dial-up connections.
Similarly, a page that looks ugly might be "crap" to a graphic designer, and a page with source code that doesn't validate might be "crap" to an HTML purist. But unless there's statistical evidence that uncompressed JPEG images and non-validating code are associated with a lack of intrinsic information value (see Google's mission statement), they're unlikely to be "signals of crap" from Google's point of view.