Re: [xml-dev] ten years later, time to repeat it?

[Elliotte Harold <elharo@metalab.unc.edu>]

Simon St.Laurent wrote:

> The lack of typing is a problem for some people.  However, I'd argue 
> that inserting typing into the core of XML would create much larger 
> problems that are harder to fix than just using xsi:type is for people 
> who want such typing.  (And xml:type is a feature addition that's 
> incompatible with subsetting in any event.)

Then they can add it using their own mechanisms. That's fine. The X 
stands for eXtensible. But we shouldn't bake one type system into the 
core to the exclusion of all other valid and useful type systems.

> I've argued for years that people who want strong typing in their data 
> transfer would be wiser to create a different format that addresses 
> their needs instead of forcing their needs into a format that wasn't 
> really designed to do what they want.

Ultimately that would founder for the same reasons types in XML core are 
a bad idea. Syntax is interoperable. Semantics aren't. Everyone needs a 
different type system, sometimes a wildly different type system. 
Encoding it in XML or another format doesn't change that.

There most certainly is a place for non-XML formats, but that's not 
going to magically make everyone agree on what a date really is or 
represents.

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