Date: May 7, 1999
Present: Anastasia Cheetham, Lake Porter, David Bolter, Jutta Treviranus,
Tamar Weiss, Joseph Scheuhammer.
Regrets: Karen McCall.
We discussed the purpose of these meetings. The first task was to review, discuss, and suggest modifications to the non-speech audio used in the audio look and feel. The second task is to review the sound sequences, including both speech and non-speech, for each swing component. Upon review, we have successfully reviewed all the non-speech sounds What remains is:
Jutta made a point that anything closing should have a "hard" end. This is meant to communicate a sense of finality. Upon reflection it is not completely clear what "anything closing means". A window for sure, but how about a menu. Does it close when it is dismissed? Note that currently no sound is made when a menu "closes" per se. Usually a menu is dismissed upon selection of one of its items and that item emits a sound to indicate its selection. There probably should be a close menu sound when a menu is simply "escaped" without selecting any of its items. In any case, can you think of other situations that consititute the "closing" of a gui widget?
Lake suggested that a next step is to create sets of non-speech audio centred around "themes". He compared this to Windows and Macintosh sound themes. It was generally thought that this was a good idea, but that we should attempt to have functional themes that made sense, and not do "ear candy".
Lake also wanted to know if we could simply use midi audio. Joseph will ask Allen if there is any technical problem with doing this vis-a-vis the Java Media Framework.
We listened to Lake's "pachenko" sound for identifying the tree gui element. I have nothing in my notes regarding what we thought of it, nor can I remember. Does anyone else remember?
Jutta recommended that everyone consult a web site that has any number of papers and/or discussions on audio interfaces. Joseph is to send everyone the URL for this site. This has been done; for posterity here is the URL again: http://www.public.iastate.edu/~CYBERSTACKS/Wave.htm. We should browse this material and compare it against what we have done. We should also, in a similar spirit, look into auditory cascading style sheets and/or XSL (I believe one can get to this information from the same URL).
There was a discussion regarding whether we should provide a specific audio look and feel for common dialogues over and above what their individual gui elements provide. A "file-open" or "file-save" dialogue are examples. The "debate" was whether such was actually necessary: whether the audio interface for the individual elements was enough. Since the meeting, I recall that we did treat simple "alert" type dialogues as special cases. These include Message, Warning, Confirmation, and Input dialogues. So, if we did it for alert dialogues, why not file choosers (and colour choosers) as well?
This discussion led to one concerning the difficulty in doing a dialogue-specific audio interface. The main difficulty is to override the audio feedback that is already there for the individual elements, and give a higher level presentation. This is sometimes referred to as the choreography problem by Jutta. This issue reminded me that the accessibility team at JavaSoft are developing a new accessibility interface that addresses this issue; it represents an attempt to make explicit functional relationships among the buttons, checkboxes, lists, etc. that appear in a dialogue. JavaSoft has emailed the preliminary specs to us (Jutta and Joseph) and asked for commentary. Note that they have also sent the specs to other adaptive technology centres for comments. We have so far ignored their requests. It was decided at the meeting that David (and Joseph?) should now review what they have sent us, and report on it by next meeting. There is also an intent to formulate a response to send back to them.
Jutta and Katharyn (and any other UI design person who is interested) will go through the current audio look and feel "demo" that is on the CSUN lap-top and evaluate it. This is planned for approximately May 20.
The next audio plaf meeting is scheduled, tentatively, for Thu May 27th. It was felt that this meeting should occur after the Jutta/Katharyn evaluation.
Joseph Scheuhammer 99 May 20.