Adds documentation to Pendulum.urdf.

[liangfok]

This change is 

[liangfok]

+@amcastro-tri for feature review please. Thanks!

---

Review status: 0 of 1 files reviewed at latest revision, all discussions resolved.

---

Comments from Reviewable

[RussTedrake]

i don't think that the first part of the documentation adds useful information, and it adds redundancy. i worry that it is doomed to get out of sync with the source below. for example, we typically don't document the value of constants.

[liangfok]

I see. Here is the thread that served as the original motivation for creating this PR:

https://reviewable.io/reviews/robotlocomotion/drake/4958#-KcruAXuE_LqGfb50JVE

Maybe now that @amcastro-tri agrees that the COM is at Z=-0.5, we don't necessarily need to document it?

---

Review status: 0 of 1 files reviewed at latest revision, all discussions resolved.

---

Comments from Reviewable

[amcastro-tri]

I don't think that the first part of the documentation adds useful information, and it adds redundancy.

I need to agree here. The model is so simple that a quick look tells you what's in there. The confusion was because I apparently didn't care to read the URDF specification carefully. Now that I know exactly what the <origin> tag within a <joint> actually specifies it is very easy for me to tell what this urdf describes.

The second part describing a limitation of this urdf when computing the center of mass with RigidBodyTree::centerOfMass() will undoubtedly go out of sync.

Up to you @liangfok. If you really feel strong about documenting this urdf I would recommend:

1. Remove this "problem" with "centerOfMass()". It will go out of sync.
2. Make the description more accurate. With that alone I cannot reproduce in my head what the setup is about. For instance, in what plane is the pendulum swinging? , where is it located?

---

Review status: 0 of 1 files reviewed at latest revision, all discussions resolved.

---

Comments from Reviewable

[liangfok]

Alright, since the the confusion about what the URDF describes is resolved and questions were raised about the usefulness of both the first and second half of the proposed documentation, I will close this PR.