[Edit: deleted parent asked: "May I ask why you include the word \"temporarily\" regarding prior art? ... Am I wrong? Does prior art now somehow not invalidate a claim/patent?"]
Usually a patent applicant can claim a narrow distinction between his claim and existing prior art. Consider gp's list of available distinctions from the Minority Report zoom feature; it's a litany of distinctions without any actual difference as far as the actual nature of the 'invention.'
It's a standard technique in patent law to apply for a series of patents around the invention of a competitor. The patent office can be persuaded to issue new monopolies over every kind of meaningless distinction. You can make it impossible for that competitor to make his invention by blocking every kind of trivial and obvious change and improvement in design, technique, and build. Then he will have to cross license in order to continue in business and you gain access to his patents. It's one of the ways big businesses continue to operate in the face of a broken patent system.
Apple recently overcame a post-grant rejection of their bounce-back patent on appeal. They argued that the prior art showed the technique to re-center the center of a fixed-size document so that it fit on the screen while their technique was completely distinct. Apple instead aligned the edge of a fixed size document to the edge of the screen so that it fit. The logical and geometric and mathematical and programming equivalency of those two did not overcome the PTO's affection for rhetorical distinctions and the patent was re-issued.
http://www.groklaw.net/article.php?story=20130509083407920
(Note that Samsung is very unlikely to get the re-trial in the story; the PTO's position on distinctions is widely approved in the bar.)
So, no. Prior art does not invalidate a claim in most circumstances.
In the most recent "When Patents Attack" episode, This American Life found 5,000 patents that covered the exact same 'invention.' That invention was online backup. The one 1990s patent under litigation could not be invalidated by the five thousand others between it and IBM's first implementations in the 1950s. (They eventually invalidated it another way, after trolls collected hundreds of millions of dollars.)