Simplified bedic format has been introduced to overcome two shortcomings of the original full bedic format: to get rid of binary characters (such as 0 bytes), which are difficult to generate in scripts, and to make editing dictionaries in a text editor as easy as possible. The drawback of the new format is that the dictionaries in the simplified bedic format are not readable by the bedic library or any bedic dictionary reader, and must be first processed with mkbedic command.
The following is an example dictionary in the simplified bedic format:
nid=Example dictionary
Keyword 1
{s}{ss}This is a description of the {hw/},
which can span several lines{/ss}{s}
Keyword 2
{s}{ss}A single keyword can contain multiple
senses and subsenses{/ss}
{ss}This is the second subsense of the first sense{/ss}{/s}
{s}{ss}And this is the second sense{/ss}{/s}
bedic header with the properties can be stored in a separate file, which can be specified with this option. If <infile> already contains some properties, they will be overwriten (if names match) with those in the <file>.
Specify this option if the <infile> does not contain a header (starts with dictionary entries).
id property in the header will be set to the argument of this option. This option overwrites both properties from the <infile> and the <file> specified with --header-file option.
mkbedic will print out the entries that are not discernable using current character precedence settings and therefore they are not accessible by bedic (bedic requires that all key-words are unique). Do not use search-ignore-chars and set the proper char-precedence property to avoid duplicate entries.
Currectly bedic can handle dictionaries up to about 2,147,483,647 bytes.
mkbedic may not handle very large files, since it uses libc IO functions.
--generate-char-precedence option is not supported. Run xerox on the <outfile> to find to generate a char precedence line.