A question from here: "Why would Kepler rotate towards Earth instead of having an antenna always directed to Earth"?
Kepler has a parabolic dish antenna fixed to its body.
There are many possible designs of transmitting devices. For spacecraft, parabolic dish antennas are preferred, because they have the least attenuation for radio waves with distance. But dish antennas must be pointed to the receiver.
The Kepler telescope was over-budget for a Discovery-class mission, so some compromises were accepted. One of them was to use a fixed dish instead of a dish on a steerable arm.
Instead of this:
it turned to this:
You can see the fixed dish at the lower part of the spacecraft.
From the NASA FAQ about Kepler:
D6. Why is the high-gain antenna fixed on the spacecraft rather than on a gimbaled arm?
Early in the spacecraft design phase, a decision was made to mount the antenna directly on the spacecraft. There were two reasons: it reduced the mission design and construction cost, and it reduced risk of failure of the arm to extend the antenna after launch.
Because of this, Kepler needs to interrupt its observations for transmission of collected data. It turns the whole body to point the fixed dish antenna towards Earth.
To address the comment:
why is Kepler not design in a way that it always points to Earth with the communication device simultaneously being able to collect data
Kepler is on an Earth-trailing orbit with a period of 371 days. You can see it in this picture:
Kepler's distance to Earth is changing with time as well as direction. Kepler was designed to stare at a fixed area in the sky for at least three years. Remember also that both Kepler and Earth are moving around the Sun, but stars do not. As a result, the angle between two lines (Kepler-Earth and Kepler-area of interest) is always changing. So it's impossible to set a fixed dish to always point towards Earth.
The same is for the K2 mission where Kepler stares at a fixed area in the sky for about 80 days and then switches to another area. (The areas should be close to the ecliptic because of solar light pressure balance requirements). These pictures show K2 areas as well as the initial (primary mission) area:
Why would Kepler rotate towards Earth instead having an antenna always directed to Earth?
There was no need to always have an antenna pointed directly at the Earth because communication with the Earth was infrequent, less than eight hours per month.
There are a lot more deep space assets than there are receivers to communicate with those assets. The solution is to have intermittent communications, with the space assets recording data for subsequent high speed playback during their short communication time slots with the Deep Space Network.
Time on the Deep Space Network is rather expensive, ranging from about \$1000 (US) to \$5000 per hour, plus an extra hour each of setup and teardown time. Because the Earth rotates, a full time, 24/7/365 contact would require switching between ground stations, requiring over 30 hours of DSN contact time per day, including setup and teardown time. A full-time, 24/7/365 contact would have made Kepler's DSN costs alone be well in excess of ten million dollars per year. It would also have required that Kepler have a significantly more expensive antenna than it did have.
Kepler was intended from the start to be a low cost mission. Instead of full-time contact, Kepler had once a month contacts with the DSN. For the vast majority of the month, Kepler was out of contact and was instead collecting data.