Re: [xml-dev] Ten Years Later - XML 1.0 Fifth Edition?

[Michael Champion <mc@xegesis.org>]

Norm, Jonathan, and others have a point about the process issue, but ultimately this is a decision about benefits and costs not right and wrong.  The benefit may be slight to the majority of current XML users, but the 5th edition edits remove an annoyance that at least potentially plagues people who are somewhat arbitrarily restricted in their choice of name characters because XML 1.0 is aligned with an old version of Unicode.

 

The costs seem minimal to implementers, and nonexistent to non-users of the characters allowed in 1.0.5 -- the spec and implementation code is simplified, and all currently "legal" documents remain legal.(Note the the XML WG removed the misguided insistence that a processor immediately reject any version number other than 1.0, so even XML 1.1 documents now become legal 1.0).

 

The interop problems will be borne by people who used XML 1.1 or want to use the 1.0.5 feature but will have to deal with the possibility that consuming implementations haven't updated. That's serious enough to require implementers to treat parser updates as real updates rather than bug fixes (e.g., Microsoft probably won't slip this into a service pack, but will probably wait for a major library update such as ".NET 4.0", whenever that happens) .  In short, I don't seriously object to W3C treating this as a bug in the spec, but we won't treat it as a bug fix in the XML libraries.  That's a red flag, for sure, but is it a showstopper?

 

This issue is open through the next W3C AC meeting, perhaps not coincidentally to be held in Beijing.  I'm definitely keeping an open mind until I hear more about the benefits from people whose native languages are affected. The *real* benefit however is to make XML 1.1 go away, and I just don't see the practical downside of that.  The XML Core WG has been trying to untie the knots around the Unicode updates and the poorly thought out XML 1.1 spec for years, and finally just got out the sword and sliced through the mess. That increases my faith in W3C, not lowers it.

 

Elliotte Harold <elharo@metalab.unc.edu> wrote:

> I didn't like that, but this is just beyond the pale. If this goes 
> through, I suspect I will completely lose faith in the W3C as a reliable 
> and honest maintainer of standards. Frankly, if we can't rely on the 
> stability of the base specs, then I think it may be time to give up on 
> XML (and the W3C) completely. :-(