Why not introduce xml 1.0 5rd edition and xml 2.0 at the same time?

Hello,

It seems likely that the reason XML 1.1 didn't take off was that it was
breaking compatibility, yet did not introduce features interesting 
enough to be worth the breakage. I agree that a 1.2 with just names 
fixed is likely to suffer a similar fate.

If the unicode name changes were packaged with worthy improvements to a
new version of the spec, it would make adoption considerably more
likely. And there's no shortage of low-hanging fruits for XML 2.0 that
should not be very controversial:

- removing DTDs (and perhaps PI) from the specification
- allowing compact end tags ("</>" or something equivalent)
- (perhaps) building the namespace spec inside the core spec

DTD removal alone would considerably simplify the code of compliant 
XML parsers, making them more secure. compact end tags would 
probably be popular with producers of XML content and partially 
placate the XML-is-too-verbose gang, encouraging updates to 
existing parsers or writing of new ones. Then it's easy to 
attach the unicode issue on top of this (it goes in the 
same direction).

This is not incompatible with releasing XML 1.0 5th edition, at the
same time as 2.0: making 1.0 worse than its pre-5th edition state 
(5th edition seen as a self-destruct tool) would give an extra 
incentive to jump to 2.0. Same idea could also be used to keep 
the 2.0 process moving swiftly: you could threaten detonating 
the 5th edition early to motivate people.

Received on Tuesday, 18 November 2008 14:43:20 UTC