Re: QT3.1 I18N reviews [I18N-ACTION-568]

On Mon, 2016-11-14 at 19:08 +0000, Phillips, Addison wrote:
> In preparing for this note, I did look at XPath31 and XQuery31 and
> noted that these had revision logs (for example
> Are these the deltas
> from 3.0 to 3.1? Or are they different? Is there any guidance other
> than the revision logs that would help us with a review?

Thank you for responding.

Short answer is each spec has changes since going to CR (more than one
CR in some cases) and then a final "changes since First Working Draft"
or similar summary that's really changes from 3.0, because the first
working draft in each case was 3.0.

I've added public-xsl-query (the joint list for XSLT and XQuery) in
case anyone has anything to add.

I'd thought I18N had reviewed Xpath and XQuery (and supporting specs)
at last call but I'm not certain - if that's true, though, you could
for sure focus on changes since then.

The 3.1 specs were supposed to be a small quick incremental change, and
in fact there aren't really that many changes.

If I'm right about the reviews, for XPath 3.1 it's
I.1, I.2, I.3 but not I.4 (Changes introduced in prior working drafts);
those I.4 changes include the changes from 3.0 to 3.1.

A note here - maps are associative arrays or hash tables and use a
Unicode collation to determine equality of keys, which can be any node,
sequence or value.

In functions and operators 3.1, appendix F.1
has changes since 3.0, but F.3 has changes since the previous CR.

I think this spec is where you're most likely to find any problems (and
possibly pleasant surprises too). For example, the addition of the UCA-
basee collation URIs.

In XQuery 3.1 appendix K
has K.1 K.2, K.3 as changes made after last call; K.4 is changes since

The XQueryX language spec is essentally generated from XQuery, so I
think won't have any issues not in XQuery; the language is in XML but
issues there are I think all long since worked out.

Although XDM (the data model) is also moving forward (we hope), I think
it unlikely to have had changes that need review: it's not user-facing
at all, and even XPath and XSLT and XQuery users don't get to deal with
it directly.

Serialization had a major change in 3.1 compared to 3.0 - a new
"adaptive" mode whichsays what happens if e.g. you just type a query at
an XQuery database rather than output to a URI. The Adaptive mode is
likely to be optional in any specs using Serialization (as are the
other modes - e.g. creating text or JSON might not make sense in an XML
Appendix F has the details of changes, with F.1 and F.2 being since
going to CR for the first time and F.3 since 3.0. Again, you may've
seen the adaptive method before, and most of the other changes were bug
fixes or related to HTML 5 and JSON.

I hope this helps (and answers your question).


Liam R. E. Quin <>
The World Wide Web Consortium (W3C)

Received on Monday, 14 November 2016 20:20:56 UTC