Re: [xml-dev] XML 1.1 and Unicode normalization

Michael Kay scripsit:

> What it actually says is:
> 
> "A text-processing component ... MUST NOT normalize the suspect text."
> 
> And it defines:
> 
> "A text-processing component is a component that recognizes data as
> text."

Yes.  Recognizes text, not produces text.

> We can argue about what that definition means - I would say anything
> that reads octets and treats the octets as encoded characters satisfies
> the definition. Perhaps the authors intended something different, but
> then they should have written something different.

I agree: that satisfies the definition.

-- 
XQuery Blueberry DOM                            John Cowan
Entity parser dot-com                           
    Abstract schemata                           http://www.reutershealth.com
    XPointer errata                             http://www.ccil.org/~cowan
Infoset Unicode BOM                                 --Richard Tobin