I find that interesting, because you could probably classify physical location revelation attacks with other timing-based attacks for which we've had to resort

I find that interesting, because you could probably classify physical location revelation attacks with other timing-based attacks for which we've had to resort to constant-time algorithms (e.g. key protection via constant-time cryptographic operations). There's an interesting fundamental tradeoff somewhere between optimization and information. If you make things as efficient as possible, you'll probably leak information.