SYMM WG Comments on XML Events

W3C HTML Working Group:

The SYMM WG has reviewed the XML Events Last Call working draft at
http://www.w3.org/TR/2001/WD-xml-events-20011026/. 
Here are our comments.

========================================================
1. The document states:
"Note that observer = "<element-id>" and event = "<event-type>" have
identical behavior to the begin = "<element-id>.<event-type>" attribute in
SMIL EventTiming [SMIL20]."

This sentence rightly establishes the link between the XML Event syntax and
the SMIL event syntax. The quoted paragraph implies that if there is a time
model in place when used with XML Events, it must be the SMIL 2.0 time
model, and it must be interpreted the same as a SMIL 2.0 "begin" attribute.
Therefore, if the observer and event attributes are used in a language along
with the SMIL 2.0 timing model, the "default" handler behavior (since no
handler is made explicit here) is the SMIL 2.0 begin behavior.

It would be helpful for implementers of both XML Events and SMIL timing to
add some of this detail to the document to minimize confusion and the
potential for
development of conflicting implementations.

2. There are some problems with the concept of handlers in that they have no
predefined semantics. In section 2.5 the specification provides flexibility
in not
requiring languages to use any one method of specifying event handlers.
However, it does not go far enough in requiring a language to define the
semantics of the event handlers that it uses. 

This has been discussed before and SYMM has dealt with something similar
with the SMIL 2.0 "timeAction" attribute. Some guidance on the semantics of
handlers must be provided.

This could be a predefined set of semantics, such as those that we defined
in SMIL 2.0 for the timeAction. Alternatively, it would be acceptable for
XML
Events to specify, as an integration requirement, that the integrating
language must
specify the handler semantics of each element type that may be used as a
handler in that language. 

Using SMIL 2.0 timeAction as an example, this is the "intrinsic" behavior
for the elements that can be used as an action (like handler), and either
the definition of the element or the integrating language must specify the
semantics when the element is invoked as an action.

Looking at the draft, some semantics seem to be implied for both script and
anchor elements. The draft could define the behavior of these elements when
used with XML events, or require the specification which uses XML events to
define these behaviors. Note that this latter option also means that XML
Events handlers cannot be used with generic XML. Only an integrating
language that defines the behavior of elements that are used as handlers as
use XML Events.

3. In section 3, the event names are all very non-WAI compliant. At the very
least they should include the more accessible names that are used in SMIL.
XML Events must coordinate with the WAI group here.

Aaron Cohen, Chair
for the SYMM Working Group

Received on Friday, 16 November 2001 18:01:56 UTC