See also: IRC log
We're not really quorate, but let's push on anyway.
At the end of this meeting, we opted to cancel next week's telcon.
Henry: If we leave it as I've
written it, it's just a hint to remind folks that the XML spec
doesn't require much.
... If you wanted to do better, you'd have to enumerate the infoset items that are required.
... But I can imagine that people will be frustrated if they have to *both* point to this profile document *and* enumerate the items.
Vojtech: I wonder if this small sentence opens new freedom in how to interpret this spec.
Norm: I don't think so, it's just more explicit about what isn't being said.
Henry: The question is, should we
put the additional detail in this spec.
... We'd have to make section 3 normative in a way that it isn't now.
Vojtech: It reduces the number of interpretations.
Henry: I'm not sure it's reasonable.
Norm: Why not?
Henry: Here's an example: the infoset spec has three different properties having to do with prefix bindings, any two of which will do
Norm: So we could say in that case that you have to provide two of the following three properties.
Henry: Does any one know if the widely used processors support the namespace attributes property?
Norm: The Java processors based on JAXP don't.
Henry: Some of the Python one's do. Whether they should or not is an another question.
Norm: I think it would be nice to
... Henry, would you try to draft it?
Henry: Yes, but it'll take a
little while. I won't know if it's going to be straightforward
or not until I've worked my way further into it.
... I can imagine that it might wind up being a two-dimensional structure depending on what parts of what specs you support.
... And what about characters, most specs don't return individual character information items.
... As you can see, it's going to take a little while to work it out.
Norm: Ok. I think it's good that we're going to look at this.
General agreement that the editor got all of them right.
Norm: Should I cancel 10 Feb?
Norm: Ok. Next meeting is 17 Feb.