In the context of computer programming, magic is an informal term for abstraction - it is used to describe code that handles complex tasks while hiding that complexity to present a simple interface. The term is somewhat tongue-in-cheek and carries good connotations, implying that the interface simplifies an otherwise difficult or tedious task. For example, Perl's polymorphic typing and closure mechanisms are often called "magic". The term implies that the hidden complexity is at least in principle understandable, in contrast to black magic and deep magic, which describe arcane techniques that are deliberately hidden or extremely difficult to understand.
Contents |
In recent years, however, a negative interpretation of the term has been gaining popularity. In this usage, "magic" refers to procedures which make calculations based on data not clearly provided to them, by accessing other modules, memory positions or global variables that they are not supposed to (in other words, they are not referentially transparent). According to most recent software architecture models, even when using structured programming, it is usually preferred to make each function behave the same way every time the same arguments are passed to it, therefore following one of the basic principles of functional programming. When a function breaks this rule, it is often said to contain "magic".
A simplified example of negative magic is the following code in PHP:
function Magic() { global $somevariable; echo $somevariable; } $somevariable = true; Magic();
While the code above is clear and maintainable, if it is seen in a large project, it is often hard to understand where the function Magic() gets its value from. It is preferred to write that code using the following concept:
function noMagic($myvariable) { echo $myvariable; } $somevariable = true; noMagic($somevariable);
Any SV [scalar value] may be magical, that is, it has special features that a normal SV does not have.—Larry Wall, perlguts manual page, Perl 5
This definition of magic or magical can be extended to a data type, code fragment, keyword, or machine address that has properties not shared by otherwise identical objects. The magic may be documented or undocumented.
FILE) cannot be safely copied as their addresses[1] may be magic. That is, the runtime environment may place original file handles in a hard-coded address range, and not provide file handle behaviour to a user-created copy at another address. Consequently the standard library routines accept pointers to file handles, of type FILE *, instead.while(<file-handle>) assigns the line read from the file to the variable $_, and applies the defined() function in the test so that any successfully read string, even "0" or the empty string, evaluates as true. This does not happen to <file-handle> anywhere else, or to while() with any other control expression.[2]CALL statement of BBC BASIC V is magic on the system call addresses of Acorn MOS; instead of branching to ARM code at those addresses, it raises a RISC OS software interrupt equivalent to the system call.[3] The effect is to emulate Acorn MOS sufficiently for 8-bit BASIC programs not containing assembly language to run unmodified.@% control print formatting, it accepts direct assignment of ANSI printf format strings, normally a type mismatch error.[3]stdio.h header file". The C book: Featuring the ANSI C standard. The Instruction Set (2nd ed.). Wokingham, England: Addison-Wesley Publishers. p. 234. ISBN 0-201-54433-4. "It is not safe to copy these objects within the program; sometimes their addresses may be 'magic'."