On 11.08.2010, at 00:56, jaayer wrote:



---- On Tue, 10 Aug 2010 02:18:19 -0700 Norbert Hartl  wrote ----


But as I wrote in my description of the problem I would call it

coalesceCDATASections: aBoolean

or

enableCoalescing
disableCoalescing

The downside of enable/disable pairs is the need for three message (two to modify, one to test and lazily initialize) rather than two.

The functionality that is described here is better known as coalescing. And it describes better what is going. If a parser is coalescing two things will happen. CDATA sections will be read in as text nodes and then subsequent text nodes are coalesing into a single text node.

my 2 cents,

Norbert

I think "preserve" is better, if only because "coalesceCData" just implies  a joining together of CDATA sections and says nothing about their status in the DOM tree as XMLString or XMLCData nodes. Although I am not wed to it.

I think it is hard to find a word that describes completely what ist going on. And I think that common sense/common usage is also kind of an argument. I didn't start to think of myself what would be the best describing word (quite hard as non-native speaker). If you search the net then you might see (as I did) that it is quite common that this effect is described as coalescing. That was my only reason to speak up because I think its recognition is better this way.
If you know about coalescing than the state in DOM tree is pretty obvious. The nodes can coalesce only if they are of the same kind. While a cdata _is_ a text node all cdata nodes are converted to simple text nodes and then all of the text nodes coalesce into one. The state in the DOM is always that there is a single text node after coalescing.

Norbert