Spokes in tension looked the ‘obvious’ way round for me, until I checked my and Mrs’ bikes, and they’re all ‘spokes in compression’, as are most pics on the web. Thinking further, spokes in tension means the spokes are pulling the disc rim into the hub, which puts the rim into circumferential compression, in addition to the compression on the ‘incoming’ side of the caliper. Looking at the photo nickjb posted, it’s the rim that’s buckled, not the spokes, suggesting it’s the rim that’s the critical bit, at least for that design of disc.
If the spokes are in compression, they push the rim out and it’s in circumferential tension. Adding tension to the rim and compression in the spokes might be best, since as it wears the rim could lose half its thickness – and buckling resistance depends on stiffness, which is proportional to thickness to the power four.
That’s only a very simplistic analysis, the actually stress in the disc will be really complex and I’d be amazed if manufacturers didn’t do FEA on the design, including thermal stress as well as mechanical, and lab test the prototype. It’s probably feasible to make a disc with a wider rim and narrower spokes that would be intended to run with the spokes in tension, but that disc in the OP doesn’t look like that. I wouldn’t fit it anyway, as the pads have worn down the end of the spokes as well as the rim.