C.4 Lexical Syntax

The lexical syntax for the components of identifiers is as follows:

• A WORD consists of a sequence of upper- and/or lower-case letters (A,...,Z and all the ISO Latin-1 national and accented letters except for the Icelandic `eth' and `thorn'), digits (0,...,9), and primes (').
• A sequence WORDS is written with each WORD separated by a single underscore (and no spaces).
• A sequence DOT-WORDS is the same as WORDS but starts with a dot(.).
• A sequence SIGNS is written with the following characters:

  +	-	*	/	\	&	=	<	>	|	^	~
  !	?	:	.	$	@	#	[	]	{	}
  ¡	¿	×	÷	£	©	±	¶	§
  ¹	²	³	·	¢	°	¬	µ

  thus not allowing the ISO Latin-1 signs for general currency, yen, broken vertical bar, registered trade mark, masculine and feminine ordinals, left and right angle quotes, fractions, soft hyphens, acute accent, cedilla, macron, and umlaut.
• A PLACE is written as a double underscore (__).
• A NUMBER is simply a sequence of digits.

(

)

;

,

`

"

%

The following further restrictions are imposed:

• Occurrences of `{...}' and of `[...]' in an ID must be balanced. E.g., `{[__}]' and `{__]' are not allowed.
• No keyword is ever recognized as a complete TOKEN.

  A keyword may however be used together with other words in a complete WORDS sequence, or as part of a single WORD.

• The ISO Latin-1 characters for product `×', negation `¬ ', and raised dot `·' are recognized as alternatives for the terminal symbols that get displayed that way: `*', `not', and `.', respectively.
• None of the following sequences of characters is recognized as a complete TOKEN (since they all have to be recognized as terminal symbols):

        :  :?  ::=  =  =>  <=>  .  ·  |  |->  \/  /\  ¬   {  }  [  ]
  

  They may however be used together with other characters in a complete TOKEN. E.g., `==', `:=', `||', `{|' are all recognized as complete TOKENs. The single characters `<', `*', `?', `!', and `/' are also recognized as complete TOKENs, despite their use elsewhere as terminal symbols.

  [CHANGED:]

• The lexical syntaxes of URL and PATH are distinguished by requiring the former to include an explicit transfer protocol prefix, e.g., `ftp:', `http:', `file:', which is forbidden in the latter. []
• It is assumed that white space always terminates a lexical symbol. The next lexical symbol recognized is as long as possible.