Meeting minutes
Inserting nodes
https://
https://
Steven: [explains the email]
Steven: Erik you were proposing document to document replacing being allowed
Erik: I haven't thought deeply about it though.
Erik: An XML document is a tree with attributes and special case of document node, and you insert into others or next to others.
… in general siblings but not always.
Steven: It seems to be about "what does the user intend?"
Erik: A replace action should be possible.
… replacing is done by delete and then insert.
Steven: I think we should be allowed to insert an attribute into an element, otherwise it is difficult to do
… especially since we now say that it is disallowed, you have to select the attribute list of a node
<ebruchez> https://
Erik: Patterns for data mutations
Steven: Why use context for setting attributes?
Erik: @context means "into". Ref creates siblings.
Steven: attributes are in a separate list from the children
Alain: I was looking at XQuery, and they have replace, replace value, and rename
Steven: Can I insert text into a document node?
Alain: Only whitespace
Steven: Is it useful then?
Erik: If you want to create any XML document, then yes.
Steven: So allow text and check that it is whitespace?
Alain: Yes
Steven: And inserting text as a sibling of the root element has also to be checked as whitespace
Alain: Yes
Alain: For DOM there used to be a text node for attributes, but not with DOM3
Steven: To insert an attribute, must you always use context?
… I would expect <insert ref="elem" origin="../blank/@attr"/> to work, and not sure what blocks it
Erik: I would like a single way of doing it
… maybe having multiple verbs and actions in XQuery is the better approach
<ebruchez> https://
Steven: My example above seems very obvious
… in intent
Steven: How about if we add to @position, an "into"
Steven: I will work on this further, and come with some proposals.
Steven: If you say <insert context="parent" ref="element" origin="somelist\/element"/>
?.. does it matter that the list, empty at first, becomes non-empty for subsequent nodes in the list.
Erik: I think ref is evaluated once
Alain: Agree
Steven: That's good
AOB
Steven: The ixml symposium was good. Videos will be posted.
… next meeting next week, an hour earlier in Europe.
[ADJOURN]