Hi, Bjoern and Anne-
Bjoern Hoehrmann wrote (on 6/21/08 6:18 PM):
* Web Applications Working Group Issue Tracker wrote:
Simon Pieters suggests wording similar to HTML5, in
<http://lists.w3.org/Archives/Public/public-webapi/2008Feb/0191.html>.
That is not a technically valid solution (and that particular wording
does not, in fact, apply to the core node traversal interfaces, if you
implement, say, .nextSibling as if entities had been expanded, entities
have in fact been expanded).
Anne's proposed solution is not valid either, except when applied to
DOM Core, rescinding EntityReference nodes alltogether, as the issue is
about how to implement this interface if you do have EntityReference
nodes in the tree (or want your code to work whether or not you do).
Having just researched the matter a bit more, I now have some improved
wording that is more procedural, and makes fewer assumptions and
impositions on the user agent. I believe that this wording should
satisfy Björn's concerns while producing the well-defined behavior that
Anne wants:
"EntityReference nodes are not visible to the Element Traversal
mechanism; if the tree-walker encounteres an EntityReference node, it
descends into it without informing the caller, and processes any
children as if they had been present at the place where the entity node
was found."
If this does not satisfy either of you, please respond within two weeks,
or sooner if possible.
Regards-
-Doug Schepers
W3C Team Contact, WebApps, SVG, and CDF
