Meeting minutes
Accept the minutes of the previous meeting
Accepted.
Review of open actions
Steven has no progress to report, nor has Norm.
Status reports
No status reports. (Norm thinks he did something, but can't remember what.)
New open issues
Steven: Where is the "iXML" vs. "ixml" distinction? The spec or elsewhere?
David: Both
Norm: We should try to be consistent.
David: We could offer guidance. The problem with "iXML" is we look like an Apple product.
Steven: The spec uses both; I'm pretty sure I always wrote it in lower case. I guess the mixed version has come from elsewhere.
John: We write XML in capitals, and since it's closely related, we put the "i" in the front. You wouldn't spell it "xml".
… In the grammar itself, we refer to it in lowercase always; in fact we always use lowercase in that grammar
… There's no reason to change that.
Steven: I've always written it lower case.
… I dunno. I don't plan personally to change.
Norm: I propose the answer is, no we can't get consensus on this.
David: We can't have inconsistency in the specification.
John: The media product appears to use "iXML".
Some discussion; looking at the 'other things named "ixml"' on the Invisible XML home page suggests there are a variety of choices.
Norm: Shall we put it on the mailing list?
David: Or we could leave it to the discretion of the editor.
Norm: Leave it for a week.
Issue #294, “parse tree” is not defined in the specification
Steven: We discussed this last week and didn't come to a conclusion.
Norm: I could be persuaded that it isn't unclear, but editorially the spec could be improved.
ACTION: Norm to try to clarify "parse treed" in the specification in a way that is satisfactory to everyone.
Perspectives on serialization
https://
Steven: It's a huge change throughout the whole spec.
Some discussion of what serialization means.
John: Consider indentation, that's clearly part of the serialization and not what we mean in most of the places in the spec. Serialization is turning the XML tree into a series of characters.
… I get an XML tree from the iXML process, but I look at the *serialized* characters that include entities, whitespaces, etc.
… That's not part of what we're describing in the formation of the XML from the iXML parse tree (trees/forest)
… Most of the specification is about putting those things into the tree.
… Because we're using XML almost exclusively in an XML environment, we should be very clear that serialization means XML Serialization.
Steven: So you're saying that serialization means something specific in XML and you'd prefer not to see some other use of that word in the spec.
John: Yes.
Steven: I don't know what's broken, six people produced implementations without this being a problem.
John: One of the exmaples in the tutorial didn't produce the right answer because I was indenting at serialization time!
… So I had to put in a mechanism to not indent.
(The discussion goes around in a circle for a bit.)
Norm: To your earlier point, Steven, six people most of whom came to these meetings and discussed things, knew what it meant. Why would we choose to use vocabulary that might confuse readers?
Steven: Now that I understand the problem, I'd like to have another period to read it and see if I think it's okay.
Norm: Okay, we can come back to this next week and finally resolve it.
The subraction item
Steven: We talked about that last week and there are some unrecorded actions.
… Fredrik or maybe it was John and I agreed to write some use cases.
Some discussion of the importance of use cases and working out what the limitations are wrt the implementation choices.
John: I have to give regrets for next week.
Next meeting
29 April at the usual time.
John: I'll create an issue for tracking it