Intel is a multinational corporation famous for its work with microprocessors. While widely regarded as the benchmark against which many if not all other processors are measured they have not always been flawless, particularly in the early 90’s when they shifted from naming their products with numbers to actual names. That is of course not to say that the naming process had anything to do with the fault, it’s simply a handy landmark in remembering the occurrence of the so-called Pentium flaw.
The “Pentium flaw” is an inherent calculation flaw that as built into one of the first Pentium microprocessors. It could be described as a very rare division miscalculation that caused an error in output approximately 61 times per billion calculations. It was brought into the mainstream by Thomas Nicely in October of 1994 after he had written computer programming to tally prime numbers. He noticed that rarely his calculations seemed suspect. After checking his work repeatedly and eliminating other factors he narrowed the malefactor to a defective pc from Intel. It was then verified by the company itself that the problem was already known after being discovered by a co-op student named Tom Kraljavic who was working for Intel at the time. Oddly though they took no steps at the time to correct or even directly address the problem.
While the flaw itself was extremely rare and didn’t affect casual users it could have greater consequences in scientific fields requiring more precise calculations (think chemical science, various physics experiments, and theoretical math.). Rather than come into the light and admit the error the Intel Corp addressed it vaguely and in a roundabout way until after Prof. Nicely had published a report on the flaw and was interviewed by CNN explaining the problem more thoroughly.
Caught in the act so to speak the company issued a report detailing the exact nature of the problem to be “an error [that] is only likely to occur [about] once in nine billion random floating point divides", and that "an average spreadsheet user could encounter this subtle flaw once in every 27,000 years of use." After more dissemination and public outcry (not to mention joking at the company’s expense) Intel instituted a recall and free replacement program for the effected chips in circulation. This would begin in December of the same year and extended to 1995.
In hindsight this could easily have been avoided if it had been admitted outright. Now we are familiar with how quickly things can spin out of control, especially on the internet. At the time of this blunder however it was believed that because the flaw would only occur in the scientific community that a general warning of some sort wasn’t necessary. To myself this seems more than a little ridiculous as it’s plainly obvious that the community most susceptible to the issue at hand also happens to be the one tech-savvy enough to spread the message like laser guided wild fire.
In the end the problem was handled however ham-handedly and only at the cost of about half a billion dollars and untold man-hours. If the same incident were to occur today then it would have to be handled immediately. Consumers get more than a little righteous indignation when their I-phone needs a software patch, I can only imagine the lynch mobs that would crowd the streets when those same crowds found out that the patch they just downloaded would cause a very specific error when given very particular inputs and even then when investigated to an almost obscene amount of detail. It’s hard to imagine though that a problem like that would be carried to the product phase now but that’s largely due to the hindsight and knowledge gained from this case.

Emery, V. (1996) The Pentium Chip Story: A Learning Experience, Emery. Com http://www.emery.com/1e/pentium.htm.
Nicely, T.R. (2011) (E-mail published openly on the internet), trnicely.net, http://www.trnicely.net/pentbug/pentbug.html
Janeba, M. (1995) the Pentium Problem, Willamette.edu, http://www.willamette.edu/~mjaneba/pentprob.html