i note that the xml lacks a namespace.
Yes, but AFAIK nothing prevents you from assigning this vocabulary to a
namespace if you wanted to mix it with some other vocabulary. But since
the need for such mixing doesn't arise in the document itself, why
bother with the added complexity?
could someone explain this design decision?
Well, I suppose this could have been done by having, say, one namespace for
Sieve stuff and another for annotations. But that adds a ton of complexity
with no real gains as far as I can see.
i also note that there is no separate of concerns between the editor
annotations and the substantive xml. could anyone explain this design
decision?
Because it would add unnecessary complexity to the design. The implication that
editor annotations are insubstantive is also incorrect - in a GUI those
annotations are the focus and it's the Sieve content that's secondary.
Ned