Re: SAX LexicalHandler::comment issue

[Lars Marius Garshol <larsga@garshol.priv.no>]

* Simon St Laurent
| 
| Writing a SAX filter is driving me to once again question the wisdom of
| the existence of attributes, but I can't say I'd seriously propose that
| they be discarded by default.
| 
| And the stuff you call noise would actually be easier for me to deal
| with in this context.

Perhaps so, but the difference between attributes and
comments/CDATA/entity refs etc is that the first are logically
significant, while the latter are not.

So for anyone who simply wishes to extract application information
from an XML document the presence of information about the lexical
details is simply noise. That noise could have been made very
unobtrusive with the right design, but the DOM chose to use a separate
node type for CDATA, for example, which makes things much more complex
than they need to be.

IMHO the DOM is made much more awkward than it needs to be because of
the way it handles these issues.

--Lars M.