Every patch is identified by a unique number, such as patchSG1234567, for example. The number is used in manufacturing to identify the collection of errors that the patch corrects. Typically, the reported errors that result in a patch release occur only under certain operating conditions. For this reason, installing a patch is necessary only if a system or site is experiencing a problem that the patch addresses.
The structure of a patch is similar to that of a software product (see Figure 1-1); however, the subsystems in a patch may span several products. For example, a patch for the FDDI product might contain file replacements for eoe and fddivis subsystems. Patches are different from standard products in another important way: installing a patch replaces individual subsystem files, not the entire subsystem.
Patches are not dependent on other patches; they rely only on the software components in the original version of the affected product. A patch might be incompatible with other patches, however. Compatibility requirements and installation instructions are described in installable release notes, which you can read from the distribution CD using the CDgrelnotes or CDrelnotes command. Instructions for reading release notes are included in the CD jewel-case insert.