Re: [css3-images] linear-gradient keywords and angles are opposite

On Fri, 10 Jun 2011 13:11:22 +0900, Brad Kemper <>  

> On Jun 9, 2011, at 7:33 PM, "Florian Rivoal" <> wrote:
>> I think we should replace left with leftwards, top with upwards, etc.  
>> Some people might say that leftwards is an uglier word than left, but I  
>> am sure nobody will ever be confused about what it means.
> That doesn't work out that well for corner to corner.  
> "upper-rightwards"? Or "upwards rightwards"?

I was thinking of upwards rightwards. I fully agree it is ugly, but I like  
ugly better than ambiguous. As a candidate for something that's neither,  
what do you think of Fantasai's proposal of "<point> to <point>", which  
would give "bottom to right". Or just going back at "from <point>", which  
would give "from bottom left".

> But really, the ONLY reason we actually NEED the keywords is for the  
> corner to corner cases. The single up, down, right, left cases can all  
> be written as degree directions. If not for corner to corner, you could  
> just use a single type of notation for all directions, and do away with  
> ambiguous keywords. Simplify.


> Which is why I had originally proposed sticking with a single way to  
> specify any direction (degrees), and just add a single keyword to say  
> whether or not the angle could change with the box shape the way  
> corner-to-corner gradients do

I'd need to think about that, but while I am not immediately convinced I  
like this, it fits the non-ambiguous criteria as well as the  
non-awkward-wording criteria, so maybe that could work.

> (based on a square, so 45deg would be a down-and-to-the-right  
> corner-to-corner gradient <<using the new meaning of degrees>> ).

In addition to clockwise, the WG also settled on 0 being north/up, so  
45deg would point up-and-to-the-right.

  - Florian

Received on Friday, 10 June 2011 04:55:10 UTC