• From: Elliotte Harold <elharo@m...>
• To: Len Bullard <len.bullard@u...>
• Date: Mon, 18 Feb 2008 11:13:32 -0800

Len Bullard wrote:
> Somewhere in that reply is the logical condition that some members of the
> audience aren't human. ;-)
> 
> This part of the discussion seems to have the most knots.  Can you clarify
> the positions?  Requiring a parser to explicitly ignore characters doesn't
> seem right to me but I likely misunderstand.
> 

I'm simply proposing that we stick with the white list approach of XML 
1.0 rather than the black list approach of XML 1.1. List the characters 
that are acceptable. Don't allow characters in names that are likely to 
cause severe interoperability problems.

As a best practice, I'd go even further than I would in the spec. You 
have to name your elements in a natural language spoken by all (or at 
least the plurality of) the recipients. The chance that such names are 
not supported by XML 1.0 is miniscule. The chance that these names are 
not supported by Unicode 5.0 is non-existent.

In 1997 one could plausibly argue that Cambodian, Amharic, and Burmese 
had yet to be encoded in Unicode and therefore an open approach was 
necessary. In 2007 this no longer holds water. There simply is no well 
from which significant new characters will be drawn. Possibly a *few* 
ideographs that Chinese speakers might wish to use will yet be coined 
during XML's lifetime; but not nearly enough to justify changing the 
basic design of XML.

-- 
Elliotte Rusty Harold  elharo@m...
Java I/O 2nd Edition Just Published!
http://www.cafeaulait.org/books/javaio2/
http://www.amazon.com/exec/obidos/ISBN=0596527500/ref=nosim/cafeaulaitA/

• References:
  • RE: Ten Years Later - XML 1.0 Fifth Edition?
    • From: Len Bullard <len.bullard@u...>