1.1.1 Multiply Rooted Instance

1.1.2 Temporary Data

1.1.3 Validation Issues

1.1.4 Relevant And Repeat Processing

Note that one way to think of repeat processing is as follows. The body of repeat contains a box that gets cloned multiple times for each node that we are iterating over. Think anonymous box showing up in the presentation DOM --we need a way of styling these boxes via CSS.

If we agree with the above, then we can have the anonymous box corresponding to a particular node in the nodeset inherit its style properties from the node it corresponds to. If we then have a CSS pseudo class for styling nodes with relevant-false we are done.

We need the following from CSS to do the above:

• A pseudo-element to address the anonymous box mentioned above.

• A pseudo-class to address relevant.

• A pseudo-class to address the repeat cursor i.e. index

Resolved as above, update specification with relevant clarification

1.1.5 Default event on listener action

Resolved: we decided not to default the event to be activate in the schema --even though it is usually ev:activate.

1.1.6 XForms Namespace

We agreed that our next public draft will have a finalized CR namespace.

1.1.7 Schema Issues

57.1

Resolved Yes-- idref should be upper case

57.2

Editorial: accept

57.3

accept

57.5schema for xforms

issue: schema for group is ambiguous. Resolved: implement option C in the schema for group. Also add optional label to case.

1.1.8 ANYURI And out Of Band Upload

1.1.9 Events Round Up

1.1.10 Dropping Script Element

Handler script is being defined in multiple places --XHTML 2.0. We feel it is better for XForms to drop element script and say that it comes from the host language. Further, element script will be defined in an XML Events handler document so that handler script becomes a reusable element that is used consistently by different host languages.

1.1.11 Enabling Use Of Repeat In HTML Tables

There is a need to be able to integrate repeat behavior into host languages at points where the content model of the host language does not provide the appropriate extension hooks via modularization. A good example of this is XHTML tables where the content model is already fixed. The present approach of using repeat elements inside tables will therefore fail to validate against an XHTML schema. Here is a solution that we agreed on.

Here are some clarifications to repeat that motivate this solution.

• Repeat contains a template that is repeated over nodes in the selected nodeset.

• Repeat does not define a container around the collection of repeated items.

Given the above clarifications, we agreed on the following approach:

• Create an attribute group in the XForms namespace that contain the following attributes from element repeat

  repeat-model
repeat-bind
repeat-nodeset
repeat-startindex
repeat-number

• We prefix each of these attribute names with repeat-

Now, repeat behavior can be used with elements that dont allow extending their content model by instead placing these repeat- attributes on the appropriate element.

The above should be thought purely as a syntactic transformation, i.e., there is no change to repeat processing semantics.

Further, for purposes of understanding the above as a pure syntactic transformation, a normal repeat construct should be viewed as containing an anonymous group that wraps the contents of element repeat. Thus, consider the following:

<repeat ...>
...
</repeat>

<repeat ...>
<group>...</group>
</repeat>

<group repeat-...>
...
</group>

Additionally, to accommodate this design, we need to say that when using action setindex, attribute repeat of type idref will point at the element carrying the repeat attributes. Similarly, when using function index against a repeating structure created via the repeat- attributes defined above, we use the id of that element as the argument to function index.