Structure Of Variable Naming

It seems proper to use systematic naming convension. For it, we define base class or inheritence of naming syntax.

Consider class names. For example, "mail" implies messsage/rfc822 format text. The input into fml system is a mail, outgoing one is a mail. The variable are sub class of "mail" object, so the name is unified as PREFIX_mail_ATTRIBUTE style.

mail_XXX
mail_default_XXX

use_incoming_mail_XXX
incoming_mail_XXX

use_outgoing_mail_XXX
outgoing_mail_XXX

use_report_mail_XXX
report_mail_XXX

Other example (header):

header_XXX
header_default_XXX

article_header_XXX
use_article_header_XXX

command_mail_header_XXX
use_command_mail_header_XXX

We can describe structure of class as follows:

command {
	SOMETHING_command
	admin_command
}
	
directory {
	XXX_directory
}
	
file {
	template_file
}
	
mail {
	incoming_mail
	outgoing_mail
	report_mail
}
	
message {
	reply_message
}
	
article {
	article_digest (not use digest to show digest of article)
	article_spool  (not use spool to show  spool of article)
}

Example: Standard Pattern (e.g. log.cf lock.cf)

use_VARIABLE		=	yes or no

# append _dir for directory.
VARIABLE_dir		=	STRING

# append _file for file.
VARIABLE_file		=	STRING

VARIABLE_type		=	STRING

VARIABLE_format		=	STRING

VARIABLE_format_type	=	STRING

VARIABLE_limit		=	STRING (NUMBER but STRING)

VARIABLE_upper_limit	=	STRING (NUMBER but STRING)

VARIABLE_lower_limit	=	STRING (NUMBER but STRING)

Example 2: (e.g. acl.cf)

VARIABLE_restrictions	=	reject_ATTRIBUTE1
				check_ATTRIBUTE2
				permit_XXX

ATTRIBUTE1 			=	PATTERN1
					PATTERN2
					...

ATTRIBUTE2			=	var1
					var2

Example 3: (program name prefix)

PROGRAM_VARIABLE_ATTRIBUTE