Order is the bread and butter of document-based formats. Not so for most data-based formats. XML was designed to accomodate both classes. I believe the M community are arguing over the same thing again.
It's bad practice to introduce a dependency on order where no dependency is necessary. Using XPath one usually does not care about order. One cares about structure but not order.
Dexter is probably right that XML syntax at least does not mandate order. But this is no more relevant as there is a newer standard which is the XML Information Set (Infoset). In particular
- The Element Information Item is defined as, among other things, composed of
[children] An ordered list of child information items, in document order.
This mandates that all XML Information Set processors must preserve element order. And as more or less all XML processors already do this, the question is rather theoretical. So even if the XML specification itself does not mandate order, the XML data-model specification does "fix" this.
It is an interesting question: how exactly to represent order.
It looks like your point is that order-independence should be the default and therefore a construct would be needed to represent order. It's an interesting question but it looks like the distinction between presentation and representation is so hardwired into brains that people don't want the syntax itself to be explicit about ordering. Still, not sure about this myself but want a processing layer between the human reader/writer and the document and preferably something else than notepad.
WindowsClient: New WPF Showcase Addition: Enterprise
TechNet Edge: Windows Server 2008 R2 : New Power Management Features
