Lars Martin wrote: > Producing SAX events is the only way to develop scalable XML applications > today. If you really need to have a DOM representation of your query result you > can easily translate SAX events to DOM nodes. (the Xerces DOM parser is based > on SAX events internally)
Actually, one of my original requirements was that the language be completely independent of a model (SAX or DOM), it should be up to the implementor what the underlying traversal model is. We'll be using a compressed DOM, which will be necessary, partly because it will be a lot faster than having a fully graphed/parsed DOM (which will get heavy really quickly), and also because our logic layer will be exposed via the DOM. Defining a logic layer that is exposed via callbacks wouldn't be intuitive for most application developers, and so we'd rather avoid it. In either case, I still think the choice of DOM or SAX should be left to the underlying implementation, and should not be part of the specification itself. --Tom ------------------------------------------------------------------ Post a message: mailto:[EMAIL PROTECTED] Unsubscribe: mailto:[EMAIL PROTECTED] Contact adminstrator: mailto:[EMAIL PROTECTED] Read archived messages: http://www.xmldb.org/ ------------------------------------------------------------------
