|This article needs additional citations for verification. (July 2007)|
||This article possibly contains original research. (October 2012)|
Software rot, also known as code rot, software erosion, software decay or software entropy describes the perceived "rot" which is either a slow deterioration of software performance over time or its diminishing responsiveness that will eventually lead to it becoming faulty, unusable, and/or otherwise called "legacy" and in need of upgrade. This is not a physical phenomenon: the software does not actually decay, but rather suffers from a lack of being responsive and updated with respect to the changing environment in which it resides.
Software can deteriorate in "performance" over time and become what is commonly called "legacy" as it runs and accumulates errors; this is not generally considered software rot, though it may have some of the same consequences. Usually, such a degraded state can be remedied by completely reinitializing its state (as by a complete reinstallation of all relevant software components, possibly including operating system software); this may also remedy some kinds of software rot, whereas other software rot is irreversible, as the operating environment of the software, or components of the software itself being incompatible, and hence becoming legacy.
There are several factors responsible for software rot. These include changes to the environment in which the software operates, degradation of compatibility between parts of the software itself, and the appearance of bugs in unused or rarely used code.
When changes occur in the program's environment, particularly changes which the designer of the program did not anticipate, the software may no longer operate as originally intended. For example, many early computer game designers made assumptions about processing speed of the CPU which the games were designed for. A consequence of this was that when the CPU's clock speed was used as a timer in the game, the gameplay speed would increase with that of the CPU, making the software less usable over time.
There are changes in the environment not related to the program's designer, but its users. A user could set up the system working once and have it working flawlessly for a time. But when the system stops working correctly, or the users want to access the configuration controls, they are unable to repeat that initial step because of the different context and the unavailable information (password lost, missing instructions, or simply a hard-to-manage user interface that was first configured by trial and error). Information Architect Jonas Söderström has named this concept Onceability, and defines it as "the quality in a technical system that prevents a user from restoring the system, once it has failed".
Portions of code which are not normally executed, such as document filters or interfaces designed to be used by other programs, may contain bugs that go unnoticed. With changes in user requirements and other external factors, this code may be executed later, thereby exposing the bugs and making the software appear less functional.
Rarely updated code
Normal maintenance of software and systems may also cause software rot. In particular, when a program contains multiple parts which function at arm's length from one another, failing to consider how changes to one part affect the others may introduce bugs.
In some cases, this may take the form of libraries that the software uses being changed in a way which adversely affects the software. If the old version of a library that was previously used with the software cannot be used any longer due to conflicts with other software (as in the case of DLLs—see DLL hell) or security flaws that were found in the old version, there may no longer be a viable version of a needed library for the program to use.
Software rot is usually classified as being either dormant rot or active rot.
Software that is not currently being used gradually becomes unusable as the remainder of the application changes. Changes in user requirements and the software environment also contribute to the deterioration.
Software that is being continuously modified may lose its integrity over time if proper mitigating processes are not consistently applied. However, much software requires continuous changes to meet new requirements and correct bugs, and re-engineering software each time a change is made is rarely practical. This creates what is essentially an evolution process for the program, causing it to depart from the original engineered design. As a consequence of this and a changing environment, assumptions made by the original designers may be invalidated, introducing bugs.
Developers are often encouraged[who?] to fully understand a problem before programming a solution to it, and to keep accurate and complete software documentation. In practice, however, adding new features may be prioritized over updating documentation. Without documentation, however, it is possible for specific knowledge pertaining to parts of the program to be lost. To some extent, this can be mitigated by following best current practices with regards to internal documentation and variable names.
Active software rot slows once an application is near the end of its commercial life and further development ceases. Users often learn to work around any remaining software bugs, and the behaviour of the software becomes consistent as nothing is changing.
Many seminal programs from the early days of AI research have suffered from irreparable software rot. For example, the original SHRDLU program (an early natural language understanding program) cannot be run on any modern day computer or computer simulator, as it was developed during the days when LISP and PLANNER were still in development stage, and thus uses non-standard macros and software libraries which do not exist anymore.
Suppose an administrator creates a forum using some online forum software. He or she then heavily modifies it, adding new features and options. This process requires extensive modifications to existing code and deviation from the original functionality of that software.
From here, there are several ways software rot can affect the system:
- The administrator can accidentally make changes which conflict with each other or the original software, causing the forum to behave unexpectedly or break down altogether. This leaves him in a very bad position: as he has deviated so greatly from the original code, technical support and assistance in reviving the forum will be difficult to obtain.
- A security hole may be discovered in the original forum source code, requiring a security patch. However, because the administrator has modified the code so extensively, the patch may not be directly applicable to his code, requiring the administrator to effectively rewrite the update.
- The administrator who made the modifications could vacate his position, leaving the new administrator with a convoluted and heavily modified forum that lacks full documentation. Without fully understanding the modifications, it is difficult for the new administrator to make changes without introducing conflicts and bugs. (Furthermore, documentation of the original system might no longer be available; it might have been abandoned, or, with the passage of enough time, been lost.)
Refactoring is a means of addressing the problem of software rot. It is described as the process of rewriting existing code to improve its structure without affecting its external behaviour. This includes removing dead code and rewriting sections that have been modified extensively and no longer work efficiently. Care must be taken not to change the software's external behaviour, as this could introduce incompatibilities and thereby itself contribute to software rot.