Réf. : Re: [Fwd: Re: non-rectangular images & <img> tag]

> So far, the demand hasn't been strong enough to outweigh the
> implementation cost.

Would implementation be easier/cheaper if, instead of defining the contour 
in terms of the fully transparent pixels in the image, you defined the 
non-over-writable areas of the image in a floatmap, structured exactly 
like an imagemap in terms of rectangles, circles, and polygons?
Presumably 90% of the logic is already present in current browsers.

Also, this would make it possible to float text over JPEG images, or over 
selected parts of any image. (For example, have the text overwrite the 
sky, but not the tree in the middle of the picture...


Received on Wednesday, 25 April 2007 13:23:33 UTC