Re: [xml-dev] Serialization of XDM - Use cases / Proposal

In my mind this new specification I'm proposing would neither subsume 
the existing  serialization specifications, nor be merged ino XQuery, 
XPath and XSLT.
The former because I'm proposing an entirely different set of use cases 
and implications that the existing specification does not address, but 
on the other hand it doesn't replace the use cases the existing spec 
addresses.  Hence it is not a replacement.  Its disjoint.

The later because in practice I cannot imagine the W3C cracking open the 
XQuery, XPath and XSLT specifactions to add this.
Perhaps they 'should' (??)... but I don't believe it will happen.   I'm 
not convinced myself that its appropriate anyway as there is value in 
having smaller specs to address individual issues then bigger specs that 
try to address everything.

In the mean time though I suggest there is real world value in a new XDM 
Serialization spec that covers the use cases I have put forth (and 
probably others as well).
Perhaps it should be called something else so as not to be confused.   
Naming suggestions welcome !

As for this becoming a 'real' specification, I certainly cannot do that, 
I'm not on the W3C  ( or ISO or Oasis or any other 'standards 
organizaiton') and cant afford to join myself and even if I could it 
would require the efforts of multiple people over a long time.  Maybe 
someday !  In the mean time if I can identify a real need and propose an 
informal specification and then implement it and see if I can get others 
to implement it that can have real value , IMHO.


David A. Lee
dlee@calldei.com  
http://www.calldei.com
http://www.xmlsh.org
812-482-5224



Jim Tivy wrote:
>
> Hi Folks
>
>  
>
> Yes, there is a question of scope.
>
>  
>
> Regarding scope of an XDM Serialization recommendation, would not an 
> XDM Serialization recommendation subsume the XSLT and XQuery 
> serialization specification - or should they not be the same 
> specification since XQuery,XPath and XSLT are tied to XDM?
>
>  
>
> Jim
>
> ------------------------------------------------------------------------
>
> *From:* Michael Kay [mailto:mike@saxonica.com]
> *Sent:* Saturday, September 19, 2009 2:10 AM
> *To:* 'David A. Lee'; 'Jim Tivy'
> *Cc:* xml-dev@lists.xml.org; 'XProc Dev'
> *Subject:* RE: [xml-dev] Serialization of XDM - Use cases / Proposal
>
>  
>
>
>  
>
>     This spec describes concatenation of node sequences such that they
>     are parseable as entities.
>
>     In the case you describe I think that fits well.
>     Atleast in my reading
>
>
>         *5 XML Output Method*
>
>     This describes serializing a sequence of nodes.  I think this spec
>     would fit for your use case.
>     Atleast that is my read, I admit I have a hard time following the
>     reading of this spec (not to blame the authors, I just have a hard
>     time understanding it all)
>
>
>      
>
>     I agree. The existing W3C serialization is intended for this
>     use case and is regularly used in this way. (Though perhaps the
>     HTML output method is more suitable than the XML output method.)
>     This requirement is out of scope for this project, which needs to
>     retain information present in the XDM that the existing
>     serialization methods throw away.
>
>      
>
>     Regards,
>
>     Michael Kay
>     http://www.saxonica.com/
>     http://twitter.com/michaelhkay
>
>      
>