Re: [xml] About libxml2 and XML 1.1

[Liam R E Quin <liam holoweb net>]

On Wed, 2010-07-21 at 20:35 +0200, Michael Ludwig wrote:
[...]
> * allow usage of the most weird characters on Earth in element and
>   attribute names (upgrade to latest Unicode standard version)

And also more than 60,000 characters that are not at all weird.
However, XML 1.0 5th edition also does this.


> * IBM mainframe line endings
These re defined by Unicode in fact.

> * control character policy change
The c1 controls were outlawed explicitly (they were already meaningless
as per Unicode, but people sometimes included them by mistake; the idea
was to help detect encoding errors)

> * "full normalization" to allow binary comparison of two documents
XML canonicalization is a separate spec, unrelated to XML 1.1.

In addition, you can use XML namespaces 1.1 with XML 1.1, which allows
you to undeclare namespace bindings.

> Bottom line: don't bother, XML 1.1 is probably irrelevant to you. One
> XML expert taught me in his class that XML 1.1 should not be used, and
> people shouldn't worry about it. Use XML 1.0 instead.

Please find a different XML expert :-)

In practice we (W3C) have removed most of the need for XML 1.1 by
releasing XML 1.0 5e. The correct claim is that XML 1.1 should not be
used except where necessary.

Liam

-- 
Liam Quin - XML Activity Lead, W3C, http://www.w3.org/People/Quin/