One of the FallaciousArguments. Goes like this.
-
Assume that there is a metric for assessing the quality of something (overall, or in a given context), and a straightforward metric for determining this quality rating for any given something. (Often times this assumption does not hold; but that doesn't stop anyone from practicing this fallacy)
-
The speaker has a something which has a quality score of x (higher is better).
-
Another party (a competitor, opponent, another researcher, whoever) has a solution which has a quality of y, where y<x
-
Speaker then asserts that the minimum score for "acceptability" is x; anything scoring lower than x is unsatisfactory.
Therefore,
This is a fallacy for several reasons.
-
Converting a multi-valued figure of merit (or a continuous or even qualitative one) to a binary figure of merit (pass/fail, good/bad, etc.) is not a trivial undertaking if it is to be done properly. Such a determination must consider many factors--the intended use, the intended user, what is reasonable to expect given available technology. Nor can it be made in a vacuum--the presence of competing solutions is well-known to RaiseTheBar. The pass/fail criteria proposed by the speaker is often arbitrary--and in many case is selected for the purpose of being able to declare a competing solution garbage. Failure to do so is risking a FalseDichotomy.
-
As mentioned above, even the multi-valued figure of merit may be dubious. Is SmalltalkLanguage better for writing business applications than FortranLanguage? Damn straight. Is it better than Fortran for high-performance number-crunching? Not on your life. Likewise for many other technologies. Many peddlers/advocates of a technology make the assumption that the application in which their solution shines is the only application worth considering. And this is assuming that it's possible to even determine a figure of merit--even a qualitative one--which can be observed and agreed upon.
A related fallacy is IfItsBetterThanMineItsOverDesigned--where it is asserted that any improvement over the speaker's preferred solution is superfluous. Much of the above applies to that as well.