Re: XML Serialisation (was Re: Agenda for XML Core WG telcon of 2009 October 7)

On Wed, 07 Oct 2009 14:42:05 +0200, Henry S. Thompson <ht@inf.ed.ac.uk>  
wrote:

> The issue as I understand it (and it's actually connected to the
> parsing/error recovery issue) is that there _is_ broken XML out there,
> and it mostly comes from programs, in particular from PHP scripts
> which produce feeds by jamming together various sources of XML.

Or maybe just various sources of *strings*.


> When
> the matrix and the sources have different encodings, this produces
> broken output, e.g. [bother, can't find the thread with the examples
> of this].  So, the argument goes, we need a W3C publication which lays
> down the law on how to programmatically produce clean XML.

While I think it would be good to have such a document, I think it  
wouldn't remove broken XML from the Web, nor convince consumers that do  
error recovery for XML today to remove their error recovery.

My understanding is that the argument goes that it would be good to  
specify error recovery for XML (although doing recovery should be  
optional).


> Here's a document which in many ways could stand as the first draft of
> such an XML Serialization Best Practices REC/NOTE/???:
>
>   http://hsivonen.iki.fi/producing-xml/

Agreed.

-- 
Simon Pieters
Opera Software

Received on Wednesday, 7 October 2009 13:38:34 UTC