In computer programming jargon, a heisenbug is a software bug that seems to disappear or alter its behavior when one attempts to study it. The term is a pun on the name of Werner Heisenberg, the physicist who first asserted the observer effect of quantum mechanics, which states that the act of observing a system inevitably alters its state. In electronics the traditional term is probe effect, where attaching a test probe to a device changes its behavior.
Similar terms, such as bohrbug, mandelbug,hindenbug, and schrödinbug (see the section on related terms) have been occasionally proposed for other kinds of unusual software bugs, sometimes in jest; however, unlike the term heisenbug, they are not widely known or used.[original research?]
Heisenbugs occur because common attempts to debug a program, such as inserting output statements or running it in a debugger, usually modify the code, changing the memory addresses of variables and the timing of its execution.
One common example of a heisenbug is a bug that appears when the program is compiled with an optimizing compiler, but not when the same program is compiled without optimization (as is often done for the purpose of examining it with a debugger). While debugging, values that an optimized program would normally keep in registers are often pushed to main memory. This may affect, for instance, the result of floating-point comparisons, since the value in memory may have smaller range and accuracy than the value in the register. Similarly, Heisenbugs may be caused by side-effects in test expressions used in runtime assertions in languages such as C and C++, where the test expression is not evaluated when assertions are turned off in production code using the
Other common causes of heisenbugs are using the value of a non-initialized variable (which may change its address and/or initial value during debugging), or following an invalid pointer (which may point to a different place when debugging). Debuggers also commonly provide watches or other user interfaces that cause additional source code (such as property accessors) to be executed stealthily, which can, in turn, change the state of the program.
Time can also be a factor in heisenbugs, particularly with multi-threaded applications. Executing a program under control of a debugger can change the execution timing of the program as compared to normal execution. Time-sensitive bugs such as race conditions may not occur when the program is slowed down by single-stepping source lines in the debugger. This is particularly true when the behavior involves interaction with an entity not under the control of a debugger, such as when debugging network packet processing between two machines and only one is under debugger control.
Heisenbugs can be viewed as instances of the observer effect in information technology. Frustrated programmers may humorously blame a heisenbug on the phase of the moon, or (if it has occurred only once) may explain it away as a soft error due to alpha particles or cosmic rays affecting the hardware.
A mandelbug (named after Benoît Mandelbrot's fractal) is a bug whose causes are so complex it defies repair, or makes its behavior appear chaotic or even non-deterministic. It also refers to a bug that exhibits fractal behavior by revealing more bugs (the deeper a developer goes into the code to fix it the more bugs they find).
A schrödinbug or schroedinbug (named after Erwin Schrödinger and his thought experiment) is a bug that manifests itself in running software after a programmer notices that the code should never have worked in the first place.
A higgs-bugson (named for the Higgs boson particle) is a bug that is predicted to exist based upon other observed conditions (most commonly, vaguely related log entries and anecdotal user reports) but is difficult, if not impossible, to artificially reproduce in a development or test environment. May also be used to reference a bug that is obvious in the code (mathematically proven), but which cannot be seen in execution (yet difficult or impossible to actually find in existence).
The term was also used in 1985 by Jim Gray, in a paper about software failures (and is sometimes mistakenly attributed to him because of this publication) and also in 1986 by Jonathan Clark and Zhahai Stewart on the mailing list (later Usenet news group) comp.risks.
Heisenbugs are difficult to identify and fix; often attempting to resolve them leads to further unexpected behavior. Because the problem manifests as the result of a separate, underpinning, bug, the behavior can be hard to predict and analyze during debugging. Overall the number of heisenbugs identified should decrease as a piece of software matures.
Also cited in LeBlanc, Richard J.; Robbins, Arnold D.; Event-Driven Monitoring of Distributed Programs, in Proceedings of the IEEE 5th International Conference on Distributed Computing Systems (ICDCS), IEEE Computer Society, Computer Society Press, 1985, pp. 515-522 Google Books search: