> [G. Ken Holman's explanation]

hmmm....

so doc() and docref() do essentially the same
thing, except that doc takes a scalar argument, and docref
takes a vector (node-set) argument?

are there any dictums about what sort of URI a compliant processor
must know how to handle? for example, at least "http:" and "file:"?
how about a pipe?

how does the processor signal that:
- it doesn't know how to handle the uri's scheme
- it gets an access denied error accessing the resource
- the resource does not exist
- the resource is not well-formed
- the resource has nothing in it

what dtd is the resource validated against?

suppose i have a source xml document with an xlink:
<employee>
<employer xml:link="simple" href="http://some.company.com/employer.xml" show="embed" actuate="auto">
</employee>
and stylesheet fragment like this:
<xsl:apply-templates select="employee/*"/>
would a compliant XSLT processor have to deal with the above the same as if
it were given this source:
<employee employer="http://some.company.com/employer.xml"/>
and had this stylesheet fragment:
<xsl:apply-templates select="doc(@employer)"/>
?

does the processor timeout if it can't get the resource in time?
is that a system-property? 
is a compliant processor prohibited from loading resources 
from doc/ref expressions that aren't activated, and must any
other resource exactly once?
can it load the ones it will load in parallel and in any order?

if i understand this correctly, for the sanity of xslt programmers
it is actually now necessary for XPointer and XSLT to converge, since
doc/ref would naturally use xpointer syntax in uri, while the
rest of a style sheet uses xslt expressions?

-mda



 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list