Have you found out anything more recently about whether Microsoft will be supporting Entailment in the future of Oslo (now Microsoft CTP)?

Thanks for the comment and sorry for WordPress’ inability to properly process the XML in your text.

You are right that the use of “all” in the XSD doesn’t tell whether the order of the elements is syntactically meaningful or not. It just tells me what is valid. “sequence” doesn’t make it meaningful either, it just takes away the option of varying the order so the question doesn’t even exist whether the order is meaningful. You are 100% correct in your rectification.

In fact my statement is tautological because if the order was meaningful then by definition the XSD could not fix it (otherwise you couldn’t convey the meaning in a valid inside document).

What I should have written is:

“This reminds me of how most XML applications disobey Postel’s law because they are fronted with an XSD validator or (more commonly) and XSD-based object mapper and the application’s XSD overuses xs:sequence.”

Noah

I think you’re perpetuating a misunderstanding here. Element order is significant in XML with or without a schema. Many users are tempted to believe that an XSD declaration

means that the order of and in an instance is insignificant. Not so. The above just means that both must be present, and that either order is acceptable. Nothing in the above says that the order is insignificant; it would be quite reasonable to build an application in which:

All of that said, it is quite a common idiom for Schema users to write when what they mean is “I don’t care”. It’s also true that has more flexible support for things like inheritance, at least in XSD 1.0, so some users choose it when might be also work. Then again, it’s a matter of taste whether forcing a canonical order is better or worse than allowing several different orders, in cases where the order does not matter to the application.

I had a briefing with an analyst today that was very hot on Microsoft’s dynamic IT strategy, and was insisting that a lot of the Systems Center / SML work is all being folded into the broader Oslo/M vision. MS wants everything from the business process to the rule deciding what RPM your CPU fan should be at to be modeled /captured in a holistic environment.

I mentioned that we (Elastra) do a very similar thing with RDF/OWL and our own DSL, it seemed to the analyst that MS may be reinventing the semantic web, from a practitioner-first point of view. And that they may be able to smooth over the maturity curve because they control so many of the underlying pieces.

What I don’t understand is whether the Azure team is plugged into this effort to the level I’d expect them to be.

I was going off of this sentence in section 1.1 of the MGraph object model: “edges outgoing from a node are either unlabeled or uniquely labeled, relative to their source node”.

But I see what you do in the example. You have only one edge, it just happens to go to a group of nodes. OK, that helps me. I am not sure I understand the logic of this design decision but at least I understand now how you can model basic things like this with MGraph.

Thanks!

William

Consider this example:

module foo {
Father {
Name => “Peter”,
Children {
{ Name => “Dora”, Born => “12.3.45″ },
{ Name => “Dan”, Born => “23.4.56″ }
}
}
}

Or did I missunderstand your concern?