RE: Comments on rotationAngle

[[ readonly attribute float rotationAngle 
the angle (in degrees) that the ellipse described by radiusX and radiusY is rotated clockwise about its center. The value must be in the range [0, 90).. If the ellipse described by radiusX and radiusY is circular, then rotationAngle must be 0. ]]

When I suggested adding a default value for rotationAngle, I was thinking about the value the device is to use when it is not able to provide a measured value for it. For example, for radiusX/Y, we already say "1 if no value is known. This attribute may not be available on all user agents or platforms." Thus, for rotationAngle, I would suggest "0 if no value is known. This attribute may not be available on all user agents or platforms. "

Furthermore I'm not entirely convinced that it is necessary to mandate rotationAngle=0 when the touch area is a circle. Let's say I'm emulating a touch area which is a slanted ellipse, and radiusX shrinks from 40 to 10 while radiusY is constant at 20. It would be awkward to have to change rotationAngle from whatever value it was to 0 the moment that radiusX hits 20, and switch it back to the original value right after.

I'd say rotationAngle *should* be 0 if the touch area is a circle, but other values should be acceptable as well. (After all the attribute doesn't actually have any effect when the area is a circle.)

- Cathy.


-----Original Message-----
From: public-webevents-request@w3.org [mailto:public-webevents-request@w3.org] On Behalf Of ext Matt Brubeck
Sent: Friday, April 01, 2011 12:56 PM
To: public-webevents@w3.org
Subject: Re: Comments on rotationAngle

I accidentally responded to Cathy off-list.  (I need to configure my mail reader to make this mistake harder to make.)  Our off-list messages are below, with her permission.


On 03/29/2011 02:20 PM, Cathy.Chan@nokia.com wrote:
 > If we are to stick with the radiusX/Y definition we now have,  > the range of the rotationAngle should be (-45,45] to eliminate  > duplicate representations..
 >
 > On a separate note, we also need a default value for  > rotationAngle, as we do for radiusX/Y.

I've changed the spec to address these comments:
http://dvcs.w3.org/hg/webevents/rev/2cc15ae86d1d

rotationAngle now has a range of [0, 90), and a default value of 0. This should provide a unique valid representation for any touch area.  I think this should simplify testing and interop without any significant cost to implementers.


On 04/01/2011 09:35 AM, Cathy.Chan@nokia.com wrote:
 > Another question: is there any particular reason that rotationAngle  > is positive in the *counter*-clockwise direction? As far as I can  > tell, related specs (CSS transform, SVG and the Linux doc that led  > to this issue) have positive angles in the clockwise direction. I  > know we don't *have* to be consistent with everybody and all, but  > consistency certainly makes life easier for the developer.

Oops!  I was a math major, and assumed that *everybody* knew that positive angles are counter-clockwise:
http://en.wikipedia.org/wiki/Angle#Positive_and_negative_angles

But I see that other W3C specs are "backward," so I will change the Touch Events spec to match them.  :)

Received on Monday, 4 April 2011 18:55:23 UTC