[
Lists Home |
Date Index |
Thread Index
]
By definition, any 'infoset' for which the serialization is not XML is not an
XML infoset.
Respectfully,
Walter Perry
"Tomayko, Ryan" wrote:
> > John Cowan wrote:
>
> > "The serialization of the XML Infoset is XML"
>
> I think this statement holds true for the base infoset rec, but how does
> everyone feel about infoset extensions? I've been going over the latest
> XInclude rec and noticed that the following from section 4.5 - "Creating the
> Result Infoset" could potentially be harmful to John's statement:
>
> <<
> "Each top-level included item is assigned an extension property [included]
> with the boolean value "true". Information items which were not processed as
> top-level included items will have no value for the [included] property.
> This property might be used by applications which require knowledge of where
> inclusion has been performed."
> >>
>
> So, assuming I've run a source infoset through an XInclude processor and now
> have an augmented result infoset that I wish to serialize as XML, is there any
> way to maintain the exact state of the infoset? i.e. How do I preserve
> extension properties like "included"?
>
> Or are "extension properties" to the infoset only relevant during the lifespan
> of the parse?
>
> Sorry if this has been hashed out before (maybe in regard to PSVI), but the
> statement "The serialization of the XML Infoset is XML" seems only partially
> true given the current state of the recs and drafts.
>
> Thanks,
>
> - Ryan Tomayko
|
