Annotations: Make default namespace explicit

[ChristianGruen]

Annotations: Make default namespace explicit

[ChristianGruen]

@rhdunn Thanks for your assessment.

I wanted to keep the %private and %public annotations rules unchanged (but I like your suggestion to revise the text).

The reasons for my suggestion are:

• The annotation namespace URI is the only one for which no prefix exists.
• The prefix-less syntax implies that the standard annotations have no namespace. With the prefix, you can make the prefix visible by writing %ann:public instead of %public. That’s similar to fn:true(), which some people prefer over true() (one difference certainly is that the default function namespace can be changed).

For existing users, nothing would change, and the ann prefix can always be overwritten.

[rhdunn]

Doesn't this cause an issue due to the annotation definitions being in an implicit "default annotation namespace"? I.e. %private, etc. would need to be %ann:private if the annotation namespace were made explicit, which would break XQuery.

err is given a name in the spec because it is referenced in many places in the spec. It's given a binding in XQuery 4.0 (3.1 didn't have this in the static context) to make it easier for users to write catch blocks.

The annotation namespace is used in various places in the spec, so this could be useful for that, but I think it would make the spec more confusing. Maybe something like:

A variable declaration may use annotations to specify that the variable is %private or %public (which is the default).

to:

A variable declaration may use one of the following annotations in the annotations namespace to specify that the variable is %private or %public (which is the default).

where annotations namespace is a definition link pointing to https://qt4cg.org/specifications/xquery-40/xquery-40.html#id-namespaces-and-qnames:

[Definition: The annotation namespace URI is http://www.w3.org/2012/xquery. ] It has no prefix and is reserved for use in this specification. It is currently used for annotations and option declarations that are defined by the XML Query Working Group.

This would be consistent and allow readers of the spec to know where to look to find the annotation namespace when reading e.g. the variable declaration section.

[benibela]

It is not an annotation namespace when it is used for annotations and options

And it is called http://www.w3.org/2012/xquery not http://www.w3.org/2012/annotations

This looks like it was supposed to be used for any feature that is XQuery-only.

Then the prefix should be xquery: if it gets any

[ChristianGruen]

It is not an annotation namespace when it is used for annotations and options

Good point, I missed this (although Reece quoted the relevant part of the spec). xquery would be fine as well.

[rhdunn]

I'm happy with adding an xquery (or alternatively xq) namespace for this to the static context.

[ChristianGruen]

I'm happy with adding an xquery (or alternatively xq) namespace for this to the static context.

👍 for xq