A crack refers to the means of achieving software cracking, for example a stolen serial number or a tool that performs that act of cracking. Some of these tools are called keygen, patch, or loader. A keygen is a handmade product serial number generator that often offers the ability to generate working serial numbers in your own name. A patch is a small computer program that modifies the machine code of another program. This has the advantage for a cracker to not include a large executable in a release when only a few bytes are changed. A loader modifies the startup flow of a program and does not remove the protection but circumvents it. A well-known example of a loader is a trainer used to cheat in games. Fairlight pointed out in one of their .nfo files that these type of cracks are not allowed for warez scene game releases. A nukewar has shown that the protection may not kick in at any point for it to be a valid crack.
The distribution of cracked copies is illegal in most countries. There have been lawsuits over cracking software. It might be legal to use cracked software in certain circumstances. Educational resources for reverse engineering and software cracking are, however, legal and available in the form of Crackme programs.
The first software copy protection was applied to software for the Apple II, Atari 800, and Commodore 64 computers.. Software publishers have implemented increasingly complex methods in an effort to stop unauthorized copying of software.
On the Apple II, unlike modern computers that use standardized device drivers to manage device communications, the operating system directly controlled the step motor that moves the floppy drive head, and also directly interpreted the raw data, called nibbles, read from each track to identify the data sectors. This allowed complex disk-based software copy protection, by storing data on half tracks (0, 1, 2.5, 3.5, 5, 6...), quarter tracks (0, 1, 2.25, 3.75, 5, 6...), and any combination thereof. In addition, tracks did not need to be perfect rings, but could be sectioned so that sectors could be staggered across overlapping offset tracks, the most extreme version being known as spiral tracking. It was also discovered that many floppy drives did not have a fixed upper limit to head movement, and it was sometimes possible to write an additional 36th track above the normal 35 tracks. The standard Apple II copy programs could not read such protected floppy disks, since the standard DOS assumed that all disks had a uniform 35-track, 13- or 16-sector layout. Special nibble-copy programs such as Locksmith and Copy II Plus could sometimes duplicate these disks by using a reference library of known protection methods; when protected programs were cracked they would be completely stripped of the copy protection system, and transferred onto a standard format disk that any normal Apple II copy program could read.
One of the primary routes to hacking these early copy protections was to run a program that simulates the normal CPU operation. The CPU simulator provides a number of extra features to the hacker, such as the ability to single-step through each processor instruction and to examine the CPU registers and modified memory spaces as the simulation runs (any modern disassembler/debugger can do this). The Apple II provided a built-in opcode disassembler, allowing raw memory to be decoded into CPU opcodes, and this would be utilized to examine what the copy-protection was about to do next. Generally there was little to no defense available to the copy protection system, since all its secrets are made visible through the simulation. However, because the simulation itself must run on the original CPU, in addition to the software being hacked, the simulation would often run extremely slowly even at maximum speed.