Re: [xml-dev] nextml

[Amelia A Lewis <amyzing@talsever.com>]

On Thu, 9 Dec 2010 09:31:06 -0000, Pete Cordell wrote:
> Looking at this, would it be fair to say that your main concern is 
> trying to fix namespaces, possibly not so much in the XML files 
> themselves, but their usage in things like XPath?
> 
> Other than that, you look fairly content with XML 1.0 as it is.

Mmmmm.  No, I don't think that's *quite* fair.

It's true that I don't care about making XML more like JSON.  I think 
that my focus is on fixing XML for "authors" (where "author" is broadly 
construed, and may include folks writing programs to consume or 
manipulate XML, not just folks creating XML from scratch).  To my mind, 
XML 1.0 has adequate expressiveness, but some horrible mistakes that 
make it far too difficult to produce and consume.

I do think that fixing the namespace mess is imperative.  I'll also 
reiterate that I'd like to see something that could be algorithmically 
mapped to XML 1.0 + Namespaces without loss; I think that defining that 
sort of specification is achievable in a reasonable time, and the tools 
required could be distributed separately, without requiring a wholesale 
upgrade of existing toolchains--but it's more interesting, to me, than 
a specification constrained to be no more than best practices for 1.0.

I've also reconsidered some things, based on responses.  It appears 
that the general consensus of experts in i18n is that UTF-8 and -16 are 
sufficient--would that mean refusing to accept other encodings 
altogether?  If a *single* encoding were required, then problems with 
inclusion/embedding might go away.

Also, DTD ... okay, removing them makes sense.  Remove them, and the 
problem of a doctype decl appearing as a result of embedding a document 
in another goes away (encoding issues may remain).  Also disappearing: 
notations, parameter entities--that's a lot of complexity gone like 
*snap*.  Unfortunately, general parsed entities also disappear--that's 
not so good, from my perspective (and Liam nailed the reason for it: 
XInclude can't define something that is permitted to appear in an 
attribute value; entity definitions can).  Predefining a larger group 
may help, but doesn't solve things.  Also, with no internal subset, 
security APIs may be forced to support xml:id rather than relying on 
the internal subset to identify attributes of type ID.

Amy!
-- 
Amelia A. Lewis                    amyzing {at} talsever.com
Razors pain you; Rivers are damp; 
Acids stain you; And drugs cause cramp.
Guns aren't lawful; Nooses give; 
Gas smells awful; You might as well live.
                -- Dorothy Parker, "Resume"