Jack say it is more efficient, using fewer bytes over the network, and send the input file directly without conversion to big-endian, etc. Explain with a short example if this argument is correct.?

John disagrees with Jack. He says an input file of size F bytes with N numbers in it will use F/4 bytes over the network, with the assumption that each number is between 1000-9999 inclusive. Is John correct?

Peter agrees with John but has a different formula in mind. He says each input number may be up to 100,000,000 and it will use F/4 bytes over the network. Is this correct?