Implementation question

Hi everybody,

is it on-topic to ask DOM implementation questions here or is this
mailing-list only for DOM specifications discussions?

Specifically, I'd like to understand how to go from an implementation
that handles any XML document to one that handles XML documents were
the tags have specific meaning, i.e. SVG or XHTML. Do implementations
subclass the Element class and override Document.createElement() to
create the appropriate object? Or is the DOM tree created out of the
input file and then a parallel, specialised, structure is instantiated
and appropriately kept in sync with it? If both approaches are valid,
is there one approach that is "more standard" than the other?

Manu