Re: [xml-dev] Allowable characters for NCName, assuming the ASCIIcharacter set?

[Ronald Bourret <rpbourret@rpbourret.com>]

Depends on what you mean by "allowable". If you mean, what characters 
will appear in your documents (where you presumably limit the characters 
to ASCII characters in some fashion), I believe you're correct. If you 
mean what characters an XML Schema-based validator will accept, it would 
be all of the Unicode characters that match the NCName production, 
regardless of whether they're ASCII or not.

-- Ron

Roger L. Costello wrote:

> Hi Folks,
>  
> Suppose that I create XML documents, restricting myself to just using the
> ASCII character set.
>  
> And suppose that I declare an element to have the datatype NCName:
>  
>     <element name="foo" type="NCName"/>
>  
> What are the allowable characters for <foo>?
>  
> I believe that the answer is: [a-zA-Z_][a-zA-Z0-9.-_]*
> 
> Here's how I arrived at my answer:
> 
> The production rule for NCName in the XML specification:
> 
> NCName ::=  (Letter | '_') (NCNameChar)* 
> 
> NCNameChar ::=  Letter | Digit | '.' | '-' | '_' | CombiningChar | Extender 
> 
> Given that I am just using the ASCII character set,
> 
>     Letter is a-zA-X
>     Digit is 0 - 9
>     CombiningChar and Extender are characters outside the ASCII character
> set (I think)
> 
> Do you agree that, given the restriction of using only ASCII characters, the
> set of characters that can be used in <foo> is: [a-zA-Z_][a-zA-Z0-9.-_]*