In physics, a sign convention is a choice of the physical significance of signs (plus or minus) for a set of quantities, in a case where the choice of sign is arbitrary. "Arbitrary" here means that the same physical system can be correctly described using different choices for the signs, as long as one set of definitions is used consistently. The choices made may differ between authors. Disagreement about sign conventions is a frequent source of confusion, frustration, misunderstandings, and even outright errors in scientific work. In general, a sign convention is a special case of a choice of coordinate system for the case of one dimension.
Metric signature 
In relativity, the metric signature could either be + − − − or − + + +. A similar dual convention is used in higher-dimensional relativistic theories. The choice of signature is given a variety of names:
+ − − −:
- Timelike convention
- Particle physics convention
- West coast convention
- Mostly minuses
- Landau-Lifshitz sign convention.
− + + +:
Regarding the choice of − + + + versus + − − −, a survey of some classic textbooks reveals that Misner, Thorne and Wheeler (MTW) chose − + + + while Weinberg chose + − − − (with the understanding that the first sign corresponds to "time"). Subsequent authors writing in particle physics have generally followed Weinberg, while authors of papers in classical gravitation and string theory have generally followed MTW (as do most Wikipedia articles related to relativistic physics). Nevertheless, the Weinberg form is consistent with Hyperbolic quaternions, a forerunner of Minkowski space.
The signature + − − − would correspond to the following metric tensor:
whereas the signature − + + + would correspond to this one:
The Ricci tensor is defined as the contraction of the Riemann tensor. Some authors use the contraction , whereas others use the alternative . Due to the symmetries of the Riemann tensor, these two definitions differ by a minus sign.
In fact the second definition of the Ricci tensor is . The sign of the Ricci tensor does not change, because the two sign conventions concern the sign of the Riemann tensor. The second definition just compensates the sign and it works together with the second definition of the Riemann tensor (see e.g. Barrett O'Neill's Semi-riemannian geometry).
Other sign conventions 
- The sign choice for time in frames of reference and proper time: + for future and − for past is universally accepted.
- The choice of in the Dirac equation.
- The sign of the electric charge, field strength tensor in gauge theories and classical electrodynamics.
- Time dependence of a positive-frequency wave (see, e.g., the electromagnetic wave equation):
- (mainly used by physicists)
- (mainly used by engineers)
- The sign for the imaginary part of permittivity (in fact dictated by the choice of sign for time-dependence)
- The signs of distances and radii of curvature of optical surfaces in optics
- The sign of work in the first law of thermodynamics.
It is often considered good form to state explicitly which sign convention is to be used at the beginning of each book or article.
See also 
- Orientation (vector space), also known as "handedness".