[R] fitting gamma using glm

[Daniel Bogod]

The data in Table 1, which is Table T.1 in Cox & Snell's Applied
Statistics, gives
the intervals in service-hours between failures of the
air-conditioning eequipment in 10 Boeing 720 jet aircraft. The
following possible models are under consideration:
(a) separate gamma distributions fitted to all aircraft, with 20 parameters;
(b) separate gamma distributions with a common shape, with 11 parameters;
(c) espearate exponential distributions to all aircraft (shape= 1,
separate μ), with 10
parameters;
(d) common exponential distribution to all aircraft (shape = 1), with
1 parameter.

http://www.utstat.toronto.edu/reid/sta410/hw2.pdf


I am having trouble in fitting gamma to the 10 different airplanes
using only one shape parameter.

Hopefully, somebody can suggest  a way to specify the model to be used in R.
I tried to join all 10 datasets into 1, but the estimate I get in
gamma.shape() is not reasonable.

Thank you