log-linearized model and steady state

[sw41]

Hi,
I have two questions here, thank you for the help.

1) If I would like to get % deviation from the steady state, should I write the model as exp (x), or exp (xx)?
Also, if one of the variable name is tauc, will it be readable if I write exp (tauctauc)?

2) If I want to get solution of % deviation from the steady state, should I take log to all the original steady state so the Dynare will evaluate the log variable around the natural log of steady state, as Stephanie Schmitt-Grohe and Martin Uribe (2004) Matlab code did?

Thank you again for the reply.

Best,
Shu-Ling