URDF/SDF snafu

When I print the Vision60, the joints are listed as below. Clearly, the name in the SDF is already a number, but we seem to invent our own, *non-matching* number :-)
```
JOINTS:
10 id=0
parent link: body
child link: hip2
pMc: 0 -0 0, 0.325 -0.1575 0
4 id=1
parent link: hip2
child link: upper2
pMc: 0 -0.5 0, -0.207198 0 -0.0599282
5 id=2
parent link: upper2
child link: lower2
pMc: 0 0.76 0, -0.023523 0 -0.096449
11 id=3
parent link: body
child link: hip3
pMc: 0 -0 0, -0.325 -0.1575 0
6 id=4
parent link: hip3
child link: upper3
pMc: 0 -0.5 0, -0.207198 0 -0.0599282
7 id=5
parent link: upper3
child link: lower3
pMc: 0 0.76 0, -0.023523 0 -0.096449
8 id=6
parent link: body
child link: hip0
pMc: 0 -0 0, 0.325 0.1575 0
0 id=7
parent link: hip0
child link: upper0
pMc: 0 -0.5 0, -0.207198 0 -0.0599282
1 id=8
parent link: upper0
child link: lower0
pMc: 0 0.76 0, -0.023523 0 -0.096449
9 id=9
parent link: body
child link: hip1
pMc: 0 -0 0, -0.325 0.1575 0
2 id=10
parent link: hip1
child link: upper1
pMc: 0 -0.5 0, -0.207198 0 -0.0599282
3 id=11
parent link: upper1
child link: lower1
pMc: 0 0.76 0, -0.023523 0 -0.096449
```

Start by reproducing the Vision60 joint listing shown in the issue and trace the URDF/SDF loading path that assigns joint IDs. Compare the names or numeric identifiers read from SDF with the IDs printed in the JOINTS output; done means the reported joint identifiers match the model without breaking the parent-child relationships.

Written by the indexing model from the issue text.