Michael Kay

>I do not see a way how $b/*[1] and $b1/*[1] return the same JNode (ditto $b/*[4] and $b2/*[1]). Firstly, it's the "same" JNode by definition if it has the same...

>But these nodes have different properties, namely they have different parents: You're clearly imagining it differently from the way I am.

I'm currently thinking along the following lines: * The child axis changes so that it decomposes a sequence of two or more items to produce two or more GNodes. *...

``` [ { 'doc': abc }, { 'doc': def } ] It would be helpful if .//text() would allow us to traverse into the XML structure, without having to care...

The confusion is for the user trying to guess what the function is going to do. If they ask to decode with offset 0 and size 40, and there's a...

>So bin:decode-string($in, $encoding, $offset, $size) should be the same as bin:decode-string(bin:part($in, $offset, $size), $encoding) The trouble with that is that if you want to decode several parts of a message...

I propose that we include a worked example showing how to exploit records and methods to implement (in effect) an AtomicSet data type. From the user perspective, this will offer:...

I believe that we now have easy-to-understand example in the spec - the "set" library for both XSLT (5.5.3 [Example: Defining an Atomic Set]) and XQuery 5.20.4 [Example: Defining an...

Thanks for the excellent comments, they help me to focus my own thoughts about the proposal. I have to confess I've been distracted by the detail. There are 40 or...

That's certainly an improvement, but there's still a lot of vague language, like the reference to "a future call" - in a functional language there is no timeline and temporal...