i'd say wide. narrow means not incorporating some characters that would
naturally fit into 'white space'. if i was parsing some text i'd
consider a non-breaking space white space and i'd expect my code to
reflect that. why would you not want your code to treat a non-breaking
space or mathematical space not as white space?
Traditionally in c NBSP was not counted as white space. See http://msdn.microsoft.com/library/default.asp?url=/library/en-us/vccelng/htm/eleme_2.asp for one reference.
This may have been accidental, as c white space properties were defined with only the 7-bit ASCII character set in mind.
But it would break current c programs if NBSP were defined as white space. Logically then, if we exclude NBSP, other "hard" spaces should also not be defined as white space.
Essentially NBSP was treated by many word processors and text editors as simply a printing character, like any other printing character, with no special "spacing" properties. It was only an imitation of a space in appearance. Undefined characters in fonts might also appear as imitiations of space in many printing systems. That did not make them white space.
Of course under Unicode specifications NBSP is expect to expand like SPACE for justification and so assumes some of the attributes of SPACE.
For compatility I think it best to not include any of the non-breaking spaces as white space.
Jim Allan
