Andy Clark wrote: > > The framework is separate from the Xerces2 implementation > of that framework. I have written (and will continue to > write) parser configurations that use the framework but > stand on their own (i.e. they don't use Xerces2 components). > In that situation, why should I have to create an extra > object just so that I can plug into the API-generators > (e.g. DOMParser, SAXParser, etc.) that use my parser > configuration?
You make it sound like Elena's proposal represents a paradigm shift, but I don't think it really is. At least not because it relies on an additional object to pass data around. We already have Attributes and Augmentations that do just that. On the other hand, passing this object through a property does seem to be inconsistent with what we currently do. So, how about some middle ground where we'd have a new object NamespaceContext that is passed in parameter to StartElement? > I think this change (for such a small performance gain) > is detrimental to the framework as a whole. And I would > be against any change that introduces implicit requirements > for users implementing the framework. I believe that this > is the "accident waiting to happen" that Joe was talking > about. Hmm... I'm not sure I understand your point here. XNI is made of a bunch of interfaces. Every one of them sets requirements for users implementing the framework. I don't see how this proposal changes anything on that front (apart from the reliance on a property which I propose to avoid). -- Arnaud Le Hors - IBM, XML Standards Strategy Group / W3C AC Rep. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
