Re: transform properties

On Feb 7, 2011, at 8:54 AM, Simon Fraser wrote:

> On Jan 18, 2011, at 6:51 pm, Vynce Montgomery wrote:
> 
>> I dunno what you guys are doing with transform
>> (a) please make rotate() positive for counter-clockwise, as it is in
>> mathematics.  This is the opposite of what the (not yet standardized)
>> prototypes in webkit, mozilla, and opera do, but it's still correct.
> 
> The clockwise direction was chosen to match SVG: <http://www.w3.org/TR/SVG/coords.html#RotationDefined>
> 
> In addition, on a web page the coordinate system is top-down, so a clockwise direction is arguably more sensible.
> 
> [Unfortunately this still conflicts with the rotation direction of the recently-proposed gradients.]

...but only if you think of the gradient direction as being a rotation. The way I view it is as a linear direction, mapped to radial numbering system in which zero is a direction pointing to the right and numbers increase counter clockwise as degrees (where a ray pointing to each number from the center of the circle indicates the linear direction). 

In theory, we could have picked some other arbitrary numbering system (inclines are often measured in percentages, but are unsuitable because they only describe 1/4 as many angles), or even letters to describe the linear direction (NNE instead of 67.5deg). Using degrees progressing clockwise with zero on the right is the most familiar way of indicating any arbitrary linear direction. It is by far the most common way of indicating gradient direction in graphics and page layout software (which also, BTW, typically have their grid coordinates zero in the upper left corner, just like CSS).

Received on Monday, 7 February 2011 17:34:16 UTC