Re: Percentage lengths in SVG 1.1 sec 7.10

I think the formula comes from the average of a ellipse:
http://math.wikia.com/wiki/Ellipsoidal_quadratic_mean_radius
Maybe you create a hypothetical ellipse with the SVG's width and height and
use the length at 45 degrees.

Rik

On Fri, Jul 20, 2012 at 11:51 AM, Alan Stearns <stearns@adobe.com> wrote:

> Hey all,
>
> There's a typo in section 7.10 in the formula for percentage lengths that
> are not heights or widths. It's just an extra parenthesis before the
> slash. So it should read:
>
> sqrt((actual-width)**2 + (actual-height)**2)/sqrt(2)
>
>
> But I'm also wondering what motivates this particular formula. I'm looking
> at how a percentage value for the radius of a circle works. In a square
> 10x10 viewport it's a percentage of 10, which makes sense. You're getting
> the length of the diagonal divided by the square root of 2, which gets you
> back to height or width.
>
> In a 20x10 rectangle, you get a percentage of 15.81. I don't understand
> how that's useful in this case. You have the length of the diagonal on top
> again, but how is the square root of 2 relevant for non-square viewports?
> Is there a non-radius use case for percentage lengths in rectangles where
> this result is useful, or was this formula chosen mainly to make a square
> viewport work as expected?
>
> Thanks,
>
> Alan
>
>
>