Re: tag name state

Anne van Kesteren <annevk@opera.com> writes:

> On Wed, 29 Feb 2012 14:34:35 +0100, David Carlisle <davidc@nag.co.uk>  
> wrote:
>> _If_ we are going to differ from HTML5 at this point I think I would go
>> further. We have a hard requirement I think that any tree have a
>> serialisation as namespace well formed XML.
>
> Is that really a hard requirement? As I understand it EXI goes beyond  
> that, the DOM certainly goes beyond that, HTML goes beyond that (and  
> provides coercion rules to get back to well-formed XML), and there's  
> probably other examples. I think the approach HTML has taken here is  
> better. E.g. earlier we discussed the Char production and that XML does  
> not allow U+0008 for instance. Consensus seemed to be that we did not want  
> to replace that with U+FFFD but instead keep it in.

A whole lot of modern XML tools begin with the assumption that the
input will be namespace well-formed. I guess I don't mind if XML-ER
can produce a more literal interpretation of the input that isn't
NSWF, but I'd like there to be defined coercion rules, as you mention
above, for getting NSWF output from it.

                                        Be seeing you,
                                          norm

-- 
Norman Walsh
Lead Engineer
MarkLogic Corporation
Phone: +1 413 624 6676
www.marklogic.com