- From: Jacek Kopecky <jacek@systinet.com>
- Date: Mon, 4 Mar 2002 17:20:10 +0100 (CET)
- To: <noah_mendelsohn@us.ibm.com>
- cc: <xml-dist-app@w3.org>
Hi all,
myself, I don't think the "root" attribute is needed any more
because the encoding rules don't explicitly allow the so-called
"independent elements on top-level of serialization" and
therefore anybody using these (effectively extending our rules)
should also be able to deal with the resulting situation. In
other cases (not using "independent elements") the "root-ness" of
an element should be readily apparent from the context.
In any case, the current plan is to add the "serialization root"
notion to the data model so we'll see how unambiguously shall we
be able to define it. 8-)
There is one possible catch for those who do use "independent
elements" - it MUST NOT be expected that a header marked as
root="false" shall not be processed by the appropriate targetted
node. Such processing migh result in unexpected behavior.
Best regards,
Jacek Kopecky
Senior Architect, Systinet (formerly Idoox)
http://www.systinet.com/
On Sat, 2 Mar 2002 noah_mendelsohn@us.ibm.com wrote:
> Out of curiousity, and following on Don's note, does the latest proposed
> text define serialization root in a way that capture the essence of Don's
> example? I think users are as confused about what a root is as about how
> to label one once they've got one. If we can't unambiguously define
> serialization root, then I'm suspicous of the whole feature (I suspect the
> feature is needed, I just want to set the bar high on having clear
> explanations. If the explanation is ambiguous or circular, the feature
> will be used inconsistently, and interop will suffer anyway.) Thanks.
>
> ------------------------------------------------------------------
> Noah Mendelsohn Voice: 1-617-693-4036
> IBM Corporation Fax: 1-617-693-8676
> One Rogers Street
> Cambridge, MA 02142
> ------------------------------------------------------------------
>
>
Received on Monday, 4 March 2002 11:20:13 UTC