W3C home > Mailing lists > Public > www-html@w3.org > August 2007

Issues with phase attribute in XML Events 2

From: John Boyer <boyerj@ca.ibm.com>
Date: Fri, 10 Aug 2007 15:09:32 -0700
To: www-html@w3.org, www-html-editor@w3.org
Cc: Forms WG (new) <public-forms@w3.org>
Message-ID: <OF57716051.1AC17E79-ON88257333.00767E0C-88257333.0079B983@ca.ibm.com>
1) The phase attribute should have an additional possible value of 
"target". 

It is odd that you can handle an event in the capture phase and you can 
handle it in the target+bubble phase, but that you cannot say you only 
want to handle the event if it is the target phase only.  Note that the 
term "target phase" actually does appear, 

I know you have a way of doing that using a target ID, but I'd like to be 
able to do it without IDs for a variety of reasons having to do with 
limitations of XML IDs at both design time (e.g. easily re-using XML 
fragments) and run-time (e.g. IDs within XForms repeats).


2) Moreover, I would like to express the point of view that the proper 
default *should* be "target".  Most frequently one wants to capture an 
event happening directly to an element, not to that element or its 
descendants.  XForms is rife with examples.  Suppose you have this:

<input ref="some/node">
   <label>This is <output ref="../label"/></label>
   <action ev:event="xforms-value-changed">
      <!-- Do some stuff -->
   </action>
</input>

If the value of "some/node" changes, then clearly we want to "Do some 
stuff".  But the output in the label also gets an xforms-value-changed 
when the value of some/label changes.  This event bubbles up to the input 
and voila, we "Do some stuff" that we should not be doing.

It is just wrong that one should have to put an ID on the input and put an 
ev:target on the action to indicate the input ID.  The above is clear, and 
if I want to capture events that are bubbling, I should have to say so.


3) Finally, I believe you should add the keyword "bubble" and perhaps have 
it mean the same thing as the current "default" keyword.  First of all, 
"default" is a horrible way to describe the default.  Second of all, I 
proposed above to change the default, and if you accept that proposal, 
then default won't be the default anymore.  As this is a major revision of 
XML Events, I believe you should be able to do this.

So you do need a new way to describe the phase currently called "default". 
 One problem with calling it "bubble" is that you do seem to be defining 
the bubble phase as excluding the target phase.  It was not previously 
clear to me that the were separate; I thought the target phase was just 
the first part of the bubble phase, but maybe that's because of how 
default was defined.

I can see at least a little utility in being able to capture an even only 
if it was targeted at a proper descendant; I can't say I have a strong use 
case to present though.  But if the bubble phase does exclude the target 
phase, then it seems like you will need a way to say target+bubble.  For 
symmetry, I'll bet you need capture+target too.  The only names that come 
to mind right now are"ascent" (target+bubble) and "descent" 
(capture+target).

John M. Boyer, Ph.D.
STSM: Lotus Forms Architect and Researcher
Chair, W3C Forms Working Group
Workplace, Portal and Collaboration Software
IBM Victoria Software Lab
E-Mail: boyerj@ca.ibm.com 

Blog: http://www.ibm.com/developerworks/blogs/page/JohnBoyer
Received on Friday, 10 August 2007 22:09:55 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 27 March 2012 18:16:10 GMT