Attributes do not parse as CDATA???

If I make this XML:

     <emma version="abc < 123"> </emma>

Expat complains that the '<' in the string is illegal. Looking at the 
scanAtts() function it is clear why. When parsing the attribute value, 
it scans for '<' and '&', and of course the closing quote. In other 
words, it is parsing the attribute value. But according to the XML 1.0 spec:

     "All attributes for which no declaration has been read
      SHOULD be treated by a non-validating processor as if
      declared CDATA"

Now, I know that SHOULD does not mean MUST, but it seems that expat 
should at least be able to consider attributes as being CDATA. Yet I can 
see no way to disable to parsing of '<' and '&'.

Am I missing something, either in the way expat works, or in the XML spec?

-- 
chris marrin                ,""$,
[email protected]          b`    $                             ,,.
                         mP     b'                            , 1$'
         ,.`           ,b`    ,`                              :$$'
      ,|`             mP    ,`                                       ,mm
    ,b"              b"   ,`            ,mm      m$$    ,m         ,`P$$
   m$`             ,b`  .` ,mm        ,'|$P   ,|"1$`  ,b$P       ,`  :$1
  b$`             ,$: :,`` |$$      ,`   $$` ,|` ,$$,,`"$$     .`    :$|
b$|            _m$`,:`    :$1   ,`     ,$Pm|`    `    :$$,..;"'     |$:
P$b,      _;b$$b$1"       |$$ ,`      ,$$"             ``'          $$
  ```"```'"    `"`         `""`        ""`                          ,P`
"As a general rule,don't solve puzzles that open portals to Hell"'