- From: Deborah Dahl <dahl@conversational-technologies.com>
- Date: Wed, 29 Jun 2011 10:18:39 -0400
- To: <www-dom@w3.org>
The Multimodal Interaction Working Group would like to offer the following comments on the DOM 3 Events Last Call Working Draft, which we discussed at our face to face meeting last week. Unfortunately we missed the deadline for LC comments because we were still wrapping up from the face to face, but we hope that you will still find our comments of value. 1. Because multimodal interaction allows the same user intent (a change, selection, or submission, for example) to be expressed through a variety of modalities, including speech, pen input, or accelerometer input, we believe it is important to consider more "semantic events". By semantic events we mean events that are based on user intention or an abstract semantic meaning, as opposed to the mechanism used to generate the event. It should be possible to indicate concepts like select, activate, undo, and value change without being tied to if it was a keydown event or a mousedown event or a speech input event or pressing a green button. While there may still be a desire for the mechanistic events that describe how user input was collected, there should be more information, and a clear separation around similarly fleshed out semantic events that represent what the user intention is or otherwise what the semantics of the event are. In addition to multimodal use cases this is also relevant to accessibility because user intent may be expressed through an assistive device rather than a traditional mouse and keyboard interface. 2. Specifically, there are some general purpose events that would be useful in a variety of contexts that are no longer part of the DOM spec (change, submit, reset, etc.). It seems like these general purpose events should be in the general DOM event specification. 3. Another specific comment is that the Composition Event is not clearly defined in the speech recognition case. While the compositionstart event describes how speech recognition would use it, the compositionupdate and compositionend are not as clearly spelled out (both what they should be set to and how often they should occur). 4. Editorial: a. The table of contents has some issues, for instance 6.1.2.3 is not in the table of contents. b. Appendix C doesn't have section numbers for subsections, either in the table of contents or the text, unlike the other appendices. c. Section 6.1.2.3 on Media Remote Controls has typesetting text ("Lorem ipsum dolor ..." etc.) still present. Best regards, Debbie Dahl
Received on Wednesday, 29 June 2011 14:19:31 UTC