In about four years of having a disk, I've never had an issue, so it's either an entirely theoretical issue - or only if you use the same rotor for a very long time (hardly the case, in this area of cycling ), but
the spiral should lean "forwards". Ie - the way they are pointing in the picture is correct.
I'm no materials engineer, but it's something to do with different strengths of the material. It's stronger when being compressed than stretched, and in the idea that most of the time a brake will be used when rolling forwards, it compresses the spiral arms into the centre, rather than pulling them away.
I've seen a couple of those floating rotors rend themselves apart under heavy load, and whilst I'm sure it's happened - I've never seen a "normal" rotor do it.
That's my very sketchy understanding of it, anyway.
Hopefully someone with some actual knowledge instead of a dodgy memory of something I never properly knew can provide a better answer. I'd be rather interested in the specifics, if anyone fancies a good, long type