Yeah, you really can't "farm" a chart... or at least, I have to wonder how you think that PP would stop the same behaviour. Using Rank score, some people will play the top map in each set once and others will play maps many times to build up the same score. That's not going to change... people are still entitled to as many tries at each map as they want to do (within the time limit), until they manage to get best PP score they can from it. Some players will do it in a single play of each, others will keep trying until they get it right. PP doesn't change that at all.
In fact, one could argue that PP would make this sort of "farming" worse... because instead of one map per set scoring (ie 18 or so scores), PP counts every map (ie ~50 odd). The pro who could get away with 18 plays before might be nickle and dimed out their position by someone who plays all 50. That's probably the greatest problem with using PP. It might be better to use only the top PP play from each set, however that removes one of the nice features of PP... that a map set with two or more top tier maps can benefit a player who does well on each of them (as opposed to everyone just playing the top tier map with the greatest score... there are a lot of underplayed gems out there because of the way rank score worked). Sure, you could just say that anyone wanting to compete on a chart seriously should play all 50 maps, pro or not... but as mentioned above by bomber34, there's a problem with getting more people to play them, so I don't think it would be a good choice to discourage people by increasing the time commitment. More needs to be done to get players to play charts, making it less trouble to play charts in the game would probably do more for the charts than changes to scoring.
One of the nice things about switching to PP, though, is that it cuts out the problems from the scoring curve for standard and CtB. I remember one chart where there was a 4:30 minute song with a really easy top diff worth 20mil... then there were some 1:30 minute songs which were much harder and together wouldn't sum to 20mil. The scoring curve resulted in weight given in the wrong areas. At the time, I was thinking that a solution might be some sort of normalization or flattening of the scores for the ranking charts. But PP would definitely be another way to solve this problem.