The alternative to this will then tend towards the extreme you see with papers like the one out of CERN that had more than 5000 co-authors. Eventually you need to set a limit and accept the fact that some people will get screwed, if not then it will become not much more than a participation trophy.