When we use “fix deform y delta_y y delta_xy …”, delta_xy will vary according to the the length of y.
Should we determine delta_xy considering the length of y deformed at that step (Otherwise, should we determine delta_xzy considering the length of of y obtained from previous step)?
How about the use of “fix deform xy delta_xy y _delta_y”? Doesn’t the order of parameters matter for this point?

Sorry for the mistype of the command.
I mean “fix deform y delta_y xy delta_xy …” and I’d like to know what length of “y” should be used to determine delta_xy.
If deformation y is applied first and the deformed length of y is used for xy, delta y should be determined as “strain xy multiplied by deformed y length”.
If the length y for delta xy is original y length before the deformation of y, delta y should be determined as " strain xy multplied by original y length".
Is the latter is a correct definition?

I'm still not sure what you are asking. There is no delta_y or
delta_xy keyword
in the fix deform comand. I assume you mean delta y and delta xy,
where y and xy are numbers you choose. The tilt factor xy is just a distance;
it doesn't depend on the Y box length. See the create_box command
for a definition of tilt factors, as used by LAMMPS.