Lists Home |
Date Index |
Elliotte Rusty Harold <firstname.lastname@example.org> wrote:
| But since most attributes either are CDATA or are treated as CDATA,
| that only works as long as there are no runs of more than one white
| space character, and no leading or trailing white space. Attribute
| value normalization by the parser rarely goes as far as XPath's
| normalize-space does.
True, and arguably normalize-space doesn't go far enough either (though
enough for XPath only). The basic distinction between CDATA and the other
declared value types has to do with (white-space delimited) tokenization.
XML 1.0 missed the boat here, in that tokenization still needs information
from the DTD. We might have dedicated the LITA delimiters (single quotes)
for this as a purely syntactic distinction, for instance.