Microdata's property ordering semantics are unclear (and perhaps unused) - can we simplify?

[w3c]

Microdata's property ordering semantics are unclear (and perhaps unused) - can we simplify?

[gkellogg]

We pushed ideas for ordering multiple property values to the vocabulary registry, and did not do ordering for schema.org.

[gkellogg]

IMO, only reasonable use for Microdata is schema.org, which currently does not order value output at all. Simplest thing is to say that multiple property values are unordered.

[chaals]

unordered works for me, but @danbri are you saying there is a real use case and requirement for specified ordering?

[danbri]

I was thinking to say something like “While it is harmless for e.g. authoring applications to make use of the enclosing documents natural ordering, publishers are cautioned that the output of microdata parsing does not preserve order unless dedicated structures (such as schema.org/ItemList or rdf:Collection) are explicitly used."

However, "the order of properties is not significant and authors must not rely on it being preserved"
works for me. Any serialization syntax by definition has an order, so maybe it needn’t be noted that using that order in editors is reasonable.