File selection
Object-oriented file management systems require that a file be selected before it can be accessed. File selection informs the operating system which file will subsequently be addressed. Successful selection of a new file causes the previous selection to become invalid. This means that only one file can be selected at any given time. Since FIDs may be freely chosen, certain limitations must be imposed on the free addressability of files. Otherwise, it could easily happen that several files with the same FID would be available in the file tree, and the operating system would then have to decide which file was meant. In order to avoid such ambiguity and thus be independent of the search algorithm used by the operating system file manager, the selection options for files are intentionally restricted. Things would be different if all the FIDs used in the file tree were unique. In this case, it would be easy to select the desired file across several directory boundaries. However, this situation is exactly what cannot always be guaranteed. Consequently, selection is only possible within certain boundaries, since otherwise unambiguous selection of the desired file cannot be assured. The MF, however, can always be selected from anywhere within the file tree, since its FID is unique within the file tree. Selecting a DF located in the first level below the MF is only possible from a DF at the same level or from the MF. Figure 5.18 shows examples of various types of allowed and prohibited selections.

Selecting directories (MF and DF)
The MF can be selected from anywhere within the file tree, either using a special selection option of the file selection command or by means of its FID (’3F00′), which only occurs once within the file tree. When the MF is selected, the selection state that exists immediately after the smart card is reset is restored, since the MF is implicitly selected by the operating system after a reset. DFs can be selected either via their FIDs or via their DF names, which contain registered and thus unique AIDs.

Explicit EF selection
There are basically two methods available for selecting EFs.With explicit selection, a specific command (SELECT FILE) is sent to the smart card before the actual access to the file takes place. This command includes a parameter holding the 2-byte FID of the file to be selected. After the file has been selected, it can be accessed by all subsequent commands.

Implicit EF selection
Implicit selection is the name given to the process in which a file is selected using a short file identifier passed as a parameter of a command that actually accesses the file. A number of restrictions apply to the use of implicit EF selection. It only works for EFs within the currently selected DF or MF. It is thus not possible to implicitly select a file across directory boundaries. In addition, implicit selection is possible only with certain access commands that allow a short file identifier to be passed as a parameter (such as READ BINARY, UPDATE BINARY, READ RECORD and UPATE RECORD). The major advantage of implicit selection is that it allows a file to be selected and accessed with a single command. This makes a SELECT FILE command unnecessary in many cases, which simplifies the sequence of commands. Due to the reduced need for communications, using implicit selection allows distinctly higher processing speeds to be achieved.

File selection using a path name
In addition to direct selection, the ISO/IEC standard allows two supplementary methods for explicit file selection using a path name. In the first method, the path from the currently selected file to the target file is passed to the operating system. The second method uses the path from the MF to the target file. Both methods are implemented in many smart card operating systems. Using these additional capabilities results in a measurable reduction in the time required to process command sequences.