Re: New FAQ: entities and NCRs

Hi Chris,

Thanks for the rewording, that's much better! Just an inline comment below.

On Mon, 04 Jul 2005 11:50:41 +0900, Chris Lilley <chris@w3.org> wrote:

> On Monday, July 4, 2005, 4:26:13 AM, Felix wrote:
>
>
>
> FS> Hi Richard,
>
> FS> Just a minor point about the FAQ:
>
> FS> "A character entity. This is a very different animal. All should be
> FS> predefined in the markup language definition."
>
> FS> Could be changed to
>
> FS> "A character entity. This is a very different animal. All should be
> FS> predefined in the markup language definition as XML DTD entity
> FS> declarations"
>
> FS> I'm not happy with that wording,
>
> Nor I. How about:
>
> "A character entity. This is a very different animal. All should be
> defined in document instance (internal DTD subset). Relying on the
> markup language definition (external DTD subset) is fragile, as fetching
> it is optional.."
>
> FS>  but I would like to express that
> FS> character entities rely on the XML DTD mechanism for general  
> entities,
> FS> which is e.g. not available in RELAX NG.
>
> The DTD mechanism is unaffected by the use of RelaxNG or W3C XML Schema
> validation.
>

It is unaffected, but I wanted to make a different point: If people define  
their markup language with RelaxNG, they need the DTD mechanism in  
addition. Not only the external DTD subset, but also the internal DTD  
subset for character entity declaration might be applicable if you use  
ONLY an RelaxNG processor. You'll need an XML DTD processor which expands  
the entity references before RelaxNG validation.

-- Felix

Received on Monday, 4 July 2005 03:00:25 UTC