From what I've read, despite the theoretical advantages of RAID (twice the throughput, twice the speed, etc.), theories fail yet again and reality kicks in, and the gain is just about nothing! You save two or three seconds off of half a minute of loading, or sure, if you're FRAPSing alot, two dedicated drives in RAID might be better, but it never was intended for desktops, and the gains are so small that it has no place in one. I was going to RAID with my new drives if it was worth it, but it's not.
I'd therefore say a drive with a higher RPM, like a Raptor, would be better. Personally, I don't think its worth roughly a little under three times the cost (this vs. this) for not even one and a half times the gain (it's mostly placebo effect versus real results), either, so I'd say none myself, but of the two, the Raptor is the only one that makes sense.
Edit: Going RAID for space is a bad idea, since you can usually (not always, but in most cases) get one larger drive cheaper than two smaller ones that equal it's space.
I edited the other part too though. Personally, I don't see the point in either. Disk benchmarks are only disk benchmarks. If the drive is defragmented enough and not over three quarters full, it's generally fast enough. If you're willing to drop the price on the bad price to performance ratio though (some "normal" SATA drives come close to Raptors in some areas on alot of tests), the Raptor is pretty fast. You won't get faster without going into SCSI or SSD territory.
raid will crash when it is poorly implemented on the motherboard, i'd prefer a h/w raid with a real dedicated card, not only you get real speed (they will have much greater throughput since they arent a chip on a motherboard with up to 500 Mb/s (500 is achieved on expensive and mobos dedicated for server) pathway on it and they still rely on software to work), stability but when you switch motherboards, the drives in the raid wont need to formatted.... sure it maycost around 200-300 aus dollars but it the long run it is worth it