HTML request for clearer XML serialization [was: Agenda for XML Core WG telcon of 2009 September 9]

> -----Original Message-----
> From: Henry S. Thompson [mailto:ht@inf.ed.ac.uk]
> Sent: Monday, 2009 September 14 5:39
> To: Grosso, Paul
> Cc: XML CORE
> Subject: Re: Agenda for XML Core WG telcon of 2009 September 9
> 
> Grosso, Paul writes:
> 
> > HTML request for clearer XML serialization
> > ------------------------------------------
> > Henry raised the issue that HTML folks think the XML
> > spec is broken because it doesn't define error recovery
> > and doesn't discuss serialization.
> >
> > Simon added his understanding of the issue at
> > http://lists.w3.org/Archives/Public/public-xml-core-wg/2009Jun/0007
> > (second half of the message) and a thread starting with a
> > reply from John ensued at
> > http://lists.w3.org/Archives/Public/public-xml-core-
> wg/2009Jun/thread.html#msg8
> >
> > Perhaps with this email beginning, Henry only needs to reply
> > to that thread to complete the following:
> >
> > ACTION to Henry:  Send email to the XML Core WG list
> > outlining the suggestion to define a serialization spec
> > including the rationale.
> 
> David Lee has done this for me:
> 
>   http://lists.xml.org/archives/xml-dev/200909/msg00072.html

I still don't get it.  

I gather from that posting that there are infosets that cannot
be serialized in XML.  Right, we knew that.

So exactly what is the issue here?

I thought it was "request for clearer XML serialization" or
"the XML spec is broken because it doesn't define error recovery
and doesn't discuss serialization."

So my answers are "errors are errors, a parser can just die if it
wants--we decided not to standardize recovery, do whatever you want
to do when parsing non-XML" and "the XML spec is not broken because
it defines parsing (starting with a serialization) not the other way
around; yes there are infosets that cannot be serialized as XML--they
aren't XML, do with them as you wish."

I'm not opposed to extending the Infoset spec to discuss serialization,
but I'm not eager to extend XML to allow us to serialize things that
are currently not serializable, and I'm not sure I understand why this
is a "BIG problem".

paul

Received on Tuesday, 15 September 2009 14:17:38 UTC