Re: [xml-dev] Towards XML 2.0

[David Carlisle <davidc@nag.co.uk>]

On 09/12/2010 13:18, Andrew Welch wrote:
> Currently you have to use a proper serialiser (because of namespaces,
> encoding, escaping etc),


No denying all of that, but (apart from namespaces) doesn't it all apply 
to json as well?

You can't just concat two json expressions together and expect it to be 
legal, and if you have a bad utf8 sequence it is bad whether or not it 
is xml or json, and in xml you have to quote < and & but in json you 
have to quote { { and "

and for most uses, most of the time you don't _have_ to use some xml 
serialising library. The XML here (mostly document sources, also used 
for code generation, not mechanical database dumps:-) are edited in 
emacs and just saved as text. Of course James Clark's nxml mode gives 
you a wrap on the knuckles if you try to save something not well formed, 
but this is still a long way from using some dedicated xml serialiation API.

Elliotte made the point that needing to quote in xml is more onerous as 
xml is typically used for text (rather than programming constructs) 
which is true, but that suggests that json (which also requires a strict 
"programmers" parsing model) isn't the right thing to be comparing xml 
with: html with it's much more forgiving parsing might be what you want 
to make it easy to type text without worrying about the parsing grammar 
rules. annevk's xml5 isn't a bad first step in that direction.

David

________________________________________________________________________
The Numerical Algorithms Group Ltd is a company registered in England
and Wales with company number 1249803. The registered office is:
Wilkinson House, Jordan Hill Road, Oxford OX2 8DR, United Kingdom.

This e-mail has been scanned for all viruses by Star. The service is
powered by MessageLabs. 
________________________________________________________________________