how can we see planets light years away?

They are not imaged, that's for sure.

What we see is the evidence of their existence.

Most of the time it's their gravitational influence on the star they orbit, or a magnitude change if a planetary system is "on edge" we see the stars magnitude change ever so slightly everytime the planet passes in front of it.

We know a lot about how stable orbits work. Johannes Kepler, and Isaac Newton came up with some pretty easy formulas.

If you know the mass of the star (which you can learn by studying its spectrum), and the orbital period of the planet (which you can tell by observing the gravitational perturbances, magnitude changes, etc) then you know it's distance from that star.

If you know how far away a planet is from its star, you have a good idea of a temperature range it has. Climate, you cannot tell. You need to know it the planet has an atmosphere, active geology, a magnetosphere, etc... none of which we can measure or detect from 20 light years.

They can't be seen.

The temperature and size, and therefore mass, of a star can be worked out from the spectrum of light and radio waves it puts out.

A planet can be detected in a number of ways. It may be because it slightly eclipses the light from the star, or it's orbit makes the star wobble slightly (as you would if you were swinging a small weight on the end of a string around your head).

The mass of the planet and distance from its star follow from how quickly it orbits and how much it causes the star to move.

From that, the type of planet can be deduced. For example, above a certain mass and the planet will be a gas giant.

Recently, technology has been developed allowing astronomers to measure the spectra not only from the star, but from the planet also. The elements that make up its atmosphere can be seen like this.

Anything much beyond this is probably what you've read in a local newspaper or internet outlet, written by someone who copied a news release and jollied it up a bit.

Trust me mate, these people aren't doing this with a high school education, and they have decades of refining and developing the methods they use. They don't want to be wrong, especially in the current climate (pun half-intended) of how science is politicised.

mostly extrasolar planets can't be seen directly. astronomers usually observe the reflex motion of the star as the planet orbits it. these changes in velocity are very small (a few meters per second), but still much easier to observe than light from the planet itself. the measurements give you the orbital period directly, and studying the star allows you to determine its mass and therefore, a lower bound for the mass of the planet.

size, temperature and climate are not determined by this method. in the case of Gl581g, it is known that it's 3-4 earth masses and orbits within the star's habitable zone. an object with that mass is probably rocky, so its size can be estimated. albedo and atmospheric composition, among other things necessary to determine the climate, can only be guessed at, but it can be said that IF it has water, it is quite likely to be liquid water, not ice or vapor. perhaps that's not as exciting as it sounded at first, but it is still something of a first in this field.

planets discovered by the transit method can be characterised a bit better. the size and mass can be determined more accurately, which gives a better idea of the planet's composition. the main components of the atmosphere can often be identified as well. unfortunately, none of the planets in the Gl581 system transit the star, so "g" almost certainly doesn't transit either.

Although there are a few extrasolar planets which have been directly photographed, most of them have not been, and were discovered in other ways. There are a couple of common techniques. One is the transiting technique, which allows us to detect planets based on the variation in brightness of their star as the planet passes in front of the star over the course of each orbit. This only works for planets with orbits aligned towards us though. Another technique is the wobble technique, where we measure the Doppler shift in light from a star and if we see a regular variation in that shift, it is probably due to the orbital effect of a planet. This works for orbits not aligned directly towards us, but is still fairly limited.

We already happen to know of certain relationships that exist between the brightnesses, colors and masses of stars. So after detecting a planet with one of the above techniques, we can use the known mass of the star and the orbital period of the planet to estimate the mass and orbital distance of the planet. These measurements allow us to take good guesses at what the planet might be like, based on what we know about planet formation.

This is all because of the recent announcement of a planet within an ideal zone of the star, isn't it?

Here, the actual article from the people who did the study explains in detail how they've done it: http://arxiv.org/abs/1009.5733

If you look at the pictures they post there -- you'll find it's all plots. They don't actually "see" planets. The only thing an astronomer can do is... count photons. On different wavelengths, sure (putting different filters in front of the sensing element, typically a CCD) but that's about it. The look for variations in brightness. They look for spectral lines to identify compositions. They take a few days or weeks collecting the data, and then spend the next weeks or months making calculations.

actually we can't ... only a handful of planets have been observed directly. scientists use indirect methods to discover the planets such as it's gravitational influence on the star , drop in intensity of the star when planets pass infront , etc. most of the methods are related to spectroscopy where the above mentioned methods are observed in the spectrum of the star.

the mass of a planet makes the star wobble and we observe the wobble of the star to tell what is the orbit and size

wel i don kno how tey find de mass and all tat junk but you can see them somewat ok wit a telescope tat the scientists use