Lists Home |
Date Index |
At 09:47 01/03/2003 -0500, Jim Ancona wrote:
>Sean McGrath wrote:
>>>>The instance is UnicodeWithAngleBrackets for sure. But an XML compliant
>>>>parser much turn this mixture into
>>Take all the start-element events, match 'em up with each corresponding
>>end-element event. They all nest
>>inside each other perfectly. Draw it. What do you see? :-)
>A tree. But _I_ made the tree, the parser didn't. It just output a stream
The tree is in there in the stream of events produced by the parser.
Parsers must pass the element structure on to the application.
The element structure is a tree rooted at the root element. That tree is
implicit in SAX/PYX and made explicit in DOM etc. Implicit
or explicit, it is always there though.
> If all I care about is constructing a list of the contents of all 'href'
> attributes anywhere in the document, I could easily do that without
> paying any attention to the overall structure of the document, and
> without ever constructing a tree.
For sure. But just because you did not build the tree does not mean that
the parser did not pass you everything you needed
to do so.