The idea that falsification is the ultimate test of a scientific hypothesis originated with Karl Popper. In his book "The Logic of Scientific Discovery" (1959), Popper argued that the only scientific theories worthy of the name are those whose supporters are willing to put their beliefs on the line by designating certain experiments or observations that could potentially disprove their pet theories.
In his classic "The Structure of Scientific Revolutions" (1970), Thomas Kuhn criticized Popper's idea of falsification as naive. Instead, he claimed, scientists normally operate within a common "paradigm," where they go about business as usual, doing their small part for science by conducting experiments that build on accepted breakthroughs from the past. Kuhn parts ways with Popper when he argues that any theory the scientific community agrees to take seriously, whether or not it's "falsifiable" in Popper's sense, is by definition scientific. In Kuhn's world view, progress happens not through falsification but through the cataclysmic paradigm shifts he labeled "scientific revolutions."
Enter Imre Lakatos. Not content with either Popper's falsificationism or Kuhn's paradigms and revolutions, Lakatos came up with his own model for how science progresses. Lakatos explained how truly vital and progressive scientific research programs could avoid being either shot down by Popper's falsifiability criterion or becoming decadent and ripe for revolutionary overthrow like one of Kuhn's paradigms.
Lakatos proposed his alternative to these flawed approached in his "Methodology of Scientific Research Programs" (1978). It consists of a "hard core" of basic, immutable theories; a more malleable "protective belt" of subsidiary hypotheses that soften the impact of contrary observations by adapting so that the program's core ideas remain intact; and, finally, a "positive heuristic" that supplies instructions for continually adapting, refining, and customizing the subsidiary hypotheses in ways that head off challenges to the core ideas.
According to Lakatos, a scientific research program will be successful only to the extent that it can anticipate and fend off potential attacks to its core beliefs, while at the same time showing that it can predict future phenomena as well as explain the past. Thus, his "Methodology of Scientific Research Programs" provides an important model for understanding how real science can progress without being derailed by the threat of either Popperian falsification, on the one hand, or a Kuhnian revolution, on the other.