I read through the new spec and it looks good. I have just a couple of
1. The first example in C.6 uses frigid="line=2,6". Unless you are
introducing a refrigerator identification system, I think that should be
It also says "There are four of them", but I think lines="2,6" is inclusive,
so that would be five lines, no? The output example seems to include the
fifth blank line that puts </pre> on its own line.
2. In C.7, the example of attribute copying, I would like to see a second
example that replaces xml:id. I presume that works even when the xpointer
is using the xml:id in the source file to fetch the content.
Regarding xml:id fixup, this feature only applies to the top-level included
element. If there are duplicate xml:ids in its descendants, then further
fixup will be required after inclusion, right?
3. This document does not address any improvements to xpointer. I still
use xsltproc for its implementation of the xpointer() scheme of xpointer,
but that scheme never became a standard. What is the status of using some
kind of xpath syntax in xpointer?