Re: [xml-dev] Updating XML

[PaulT <pault12@pacbell.net>]

> Well, PaulT was certainly right on when he said it could be more terse.  
> Of course, there are some areas where the facile approach to terseness 
> would just cause more bloat:
> 
> <xupdate:update select="/foo/bar">
>   <malatesta fn="sigismundo">
>     <manfredi fn="manfredo">
>       <gonzaga/>
>     </manfredi>
>   </malatesta>
> </xupdate:update>
> 
> is much better IMHO to
> 
> remove(/foo/bar)
> append-element(/foo, "bar")
> append-element(/foo/bar, "malatesta")
> set-attribute(/foo/bar/malatesta, "fn", "sigismundo")
> append-element(/foo/bar/malatesta, "manfredi")
> set-attribute(/foo/bar/malatesta/manfredi, "fn", "manfredo")
> append-element(/foo/bar/malatesta/manfredi, "gonzaga")

I agree that sequence of brutal setters is not a 
universal solutuion, but there could be another 
approach to bi-directional XPath, still based 
on /foo/bar/#value = 'value'. Apply the 
way of thinking that XPath has taken, just
apply it for 'writing'.
 
> Of course perhaps this is a straw man.  Paul, what else have you got in 
> your sleeve?

I can provide a terse XPath-alike dialect that would 
cover this testcase ... and many others testcases...  
and you can do the same, just put the concepts of 
current XPath on steroids ... Also, I can provide 
improved version of XPath that would be 
bi-direcional by design ( current XPath has 
not been designed for updates, because 
it was a part of XSL and XSL was not for 
updates e t.c. e t.c. )

I was wondering if somebody already did that. 

We have plenty of people, who are selling XML 
tools. Big companies. Significant R&D budgets.
Standards and Standards. 

I was wondering, maybe there is a terse and consistent 
language for quering / updating XML.

Rgds.Paul.