From UTFT.h (which is a similar library)
void printNumF(double num, byte dec, int x, int y, char divider='.', int length=0, char filler=' ');
Yes, there are 7 arguments but 3 of them are seldom used. You would normally just use
myGLCD.printNumF(1.23456, 1, x, y); //1.2 with 1 decimal places
myGLCD.printNumF(1.23456, 2, x, y); //1.23 with 2 decimal places
myGLCD.printNumF(1.23456, 3, x, y); //1.235 with 3 decimal places
myGLCD.printNumF(1.23456, 4, x, y); //1.2346 with 4 decimal places
In practice you might want to RIGHT justify the number with the x argument.
You only use divider if you want a comma e.g. 1,23
You only use length if you want to format neatly. (similar to x=RIGHT)
You only use filler if you want to pad with leading zero e.g. 0001.23
Personally, I don't like Henning Karlsen's choice of methods. But printNumF() seems a lot easier to use than your sequence.
Incidentally, Mr Pyner's example should use 3 decimal places and sensible x, y arguments
David.