Re: Attribute values

On Wed, Aug 15, 2012 at 2:55 AM, John Cowan <cowan@mercury.ccil.org> wrote:

> What is more (as I have been saying at every opportunity), HTML
> autocorrection works because we know what the semantics of the tags and
> attributes are.  In XML, we have no clue which elements are supposed to
> go where, so we can't effectively correct it without schema information.

Moreover, having the XML parser first perform schema-unaware error
correction makes it hard/impossible for a schema processor to do good
quality error handling. For example, corrections introduced by the
parser could generate errors from the schema processor, which would be
super-confusing for the user.

Draconian error handling is a bad idea but I don't buy into the WHATWG
position that it is necessary to define the precise behaviour of an
implementation on any input whatsoever.

James

Received on Wednesday, 15 August 2012 02:08:18 UTC