NAME
tm-format - Time Machine file format
DESCRIPTION
This document presents a portable, self-documenting and extensible file format. The general format is lines of text, each line consisting of one or more fields. Adjacent fields are separated by one `|' (pipe) character. Non-printable characters, the newline, `%' (percent), and `|' (pipe) characters are represented by their hexadecimal equivalent as `%NN'.
HEADER SECTION
A file in time-machine format starts with a two-line header that specifies the origin of the data. The first line contains the names of origin attributes. The second line lists the actual values of those attributes in the order of the attribute names. Examples of recommended origin attributes are:
- class
-
A string that specifies the type of data. Usually this is named after the tool that collected the information.
- host
-
A string with the unique name of the host on which the data was collected.
- start_time
-
Start of data collection (UNIX time in seconds).
- end_time
-
End of data collection (UNIX time in seconds).
DATA SECTION
The remainder of the file consists of one line with data attribute names, followed by zero or more lines, each line listing data attribute values in the order of the data attribute names.
The choice of data attribute names and of their value types is entirely up to the application. The implementor is encouraged to consistently use generic attribute names such as `host' and `user', and to consistently use UNIX time coordinates for all time information, so that one can use generic tools to correlate files from different sources without having to write application-specific software.
LICENSE
This software is distributed under the IBM Public License.
AUTHOR(S)
Wietse Venema IBM T.J. Watson Research P.O. Box 704 Yorktown Heights, NY 10598, USA