In the Lab: Doing strange things to CSS Backgrounds
So, last year you might remember me running through some experiments with using percentages to position your background-image in CSS. The short version: Percentages are slightly counter-intuitive but quite powerful — if it wasn’t for IE6-7 sucking by always minutely miscalculating them.
In the original post we stuck mostly to the more predictable and commonly-used percentages — from 0-100%. Today I thought we might look at some of the interesting things that happen if start to crank those numbers.
Now to be clear, this is mostly ‘theoretical CSS’ in a Stu Nicholls kind of way and probably isn’t stuff you’ll be finding a use for every day. Hopefully it will, at least, be handy for helping you to get a good grip on the subtleties of how percentages interact with images. With luck it might even come in handy at some point in the future.
Let’s look at a starting example, taking a background-image positioned
150% 0%. The zero is a no-brainer, and will obviously place the graphic on the top edge of the box.
For the horizontal position the browser will first locate a spot 50% to the right side of your DIV. It will then find a spot 50% to the right hand side of your image, and align those two points. This will actually have the effect of moving your image back towards (relative to the original point you located).
This means that if the image is small, that offset will also be small and the image will remain mostly, if not entirely hidden beyond the right edge of your DIV.
If the image and DIV are the same size, these offsets will cancel each other out, and the image will be positioned in the horizontal center of it’s DIV.
Howver, if the image is larger than it’s containing DIV, this offset will actually start to move to the LEFT of the containing DIV. Weird but it makes sense if you think about it.
Negative percentages work the same way, only to the left of screen.
However it’s when you set your background to tile horizontally that really interesting things start to happen (
background-repeat: repeat-x;). While the tiling effect will ensure your image will always be visible, setting large percentages means even a small resize of your browser will translate into a massive repositioning your background graphic. For instance, if your background-position is set to -500%, a 50px increase in your browser will slide your background-image 250px to the left.
Start your browser smallish, grab a corner and as you begin to resize the application up and down, you should see an animated ‘diorama’ effect happening around the limo. Each of the ‘stage layers’ (i.e. the cityscape, background crowd and foreground crowd) is given a larger negative percentage than the one before it (-100%, -250% and -550% respectively).
The limo is given a positive percentage of 150%, which drives it in and out of the right hand stage as you resize.
Kinda mesmerizing in a Mario Brothers kind of way, isn’t it?
So are ‘resize animations’ the future of the web?
Clearly not, but it does offer a few practical possibilities. I could imagine some nice ideas with site headers where a number of partially-transparent layers were overlayed and positioned with large percentages. Resizing the browser would change the way these layers interacted, perhaps making the header appear radically different at various browser resolutions. As an analog, think of the large changes a small twist can create in a kaleidoscope.
Even if nothing particularly practical ever comes of it, it was a lot of fun to fiddle with and I think I’ve got a rock solid grasp on how percentages work now. Hopefully it clears things up for a few others too.