Putting each "sentence" of code on its own line is actually the crutch that leads to low density and high line counts. As humans en masse become more comfortable parsing and writing machine inter-languages I imagine things will trend the other way.