Re: Statistically unsound sample
From the article: "Only exoplanets with orbital periods less than a hundred days were considered".
Turning the telescopes the other way around, this would mean that aliens would only find Mercury in our solar system. Not exactly a representative sample.
When the discovery technique relies upon a planet 'eclipsing' its star for us, then clearly we are going to miss a lot of larger orbit planets - plus of course, any system that's not on the 'right' orbital plane to pass between the star and us.
Hence whilst I think all the discoveries so far are wonderful, it seems a bit pre-emptive to start classifying a family tree of planets..
Yes, it's pre-emptive. But we classify what we can, and adjust the classification models as we learn more. That's kinda how things work.
Otherwise, at what point would you suggest we should we start classifying? How will we know when we have enough data that we can start work without being pre-emptive? Given the size of the galaxy, the biases inherent in every observation technique we've come up with, and the sheer difficulty in spotting objects like Sedna even when they're in our own system, it's clear to me that we're not going to be getting even close to an accurate picture of things for some time to come.
But even if we did have a chance of getting to a that point, we still shouldn't be waiting until we have a clearer picture of the real planet distribution before starting to classify them. The problem is that the picture can only become clear with the help of the classification work, so we really do have to do some classifying work up-front, even if we know it's going to be superseded.