You probably have seen the news about the Nobel Prize in Chemistry just awarded to Israeli researcher Dr. Daniel Schechtman, for his discovery and recognition of a crystalline structure which scientists previously thought couldn’t exist. What's especially interesting is that, according to this well-written article here in The Wall Street Journal , his work was initially rejected by journals; he was told to go back and study basic crystal theory; and even chemistry Nobel laureate Linus Pauling denounced his work.
This, of course, is not the first time that a completely rejected idea which runs contrary to conventional wisdom is eventually accepted and established. (You can make your own list, starting with Copernicus.) Reality is that a large part of the history of scientific and engineering advancement is defined through these types of markers, where such accepted theory is challenged and eventually overturned.
Thomas Kuhn discussed this process of paradigm shift in his classic work The Structure of Scientific Revolutions (it's OK to use that now-tired cliché here, since he actually popularized it in that work). [Note: If you want to plan the book, be warned—it has a lot of insight but the writing is so academically dense you'll need a machete to hack through it.]
There are many engineering examples as well. One which I like to cite is discussed in detail in the excellent book Apollo: The Race to the Moon , by Charles Murray and Catherine Bly Cox. The entire concept of the now-legendary lunar module/lander was not the original plan. The original approach was to have the astronauts land the third stage of the rocket on the moon, by “backing it down”, a direct-descent approach much like backing into a parking space. This same landing stage then would lift off for the return. It was all very much like depicted in science-fiction stories and films such as 1950 classic Destination Moon , see here .
NASA had gone pretty far along in planning the mission using the third-stage lander approach, but one man, John Houbolt, thought about it differently. He went through the numbers, insisted that the conventional approach would not work, and instead his lander-only approach was better.
Since the initial weight, propulsion, cost, and risk-analysis numbers made neither of the two choices a clear winner, Houbolt was told to “shut up”; though it is now routine, in-orbit docking had never been done, and it would have to be done around the moon, too; (and there was some NIH—not invented here—aspect to it, as well). He even went over the heads of some layers of management to pitch his case, and was firmly shot down.
But as the design details were worked out, engineers found the direct-descent approach just wouldn't work. Too many factors pushed against it, when specifics were refined and analyzed. Houbolt's idea was retrieved and re-invigorated, and looked increasingly attractive. Eventually, it became the only viable option. We know how the story ends: the lunar module/lander approach was both viable and extremely successful.
Have you ever been associated with a design approach or debug analysis at which everyone else scoffed, only to be eventually vindicated when you were proved right? ?