Cons: Others than the original author may upload content and get rewarded for something they didn’t produce. Usage may be faked and faking services could make a business faking popularity (a bit like selling Facebook “likes”). For apps, the author could create a system where users get a share of the popularity reward. While the apps offers nothing of value, people would simply visit it daily (using an automated script or something) to access it and reap a part of the reward.
In my opinion, while the idea is great in theory, any such system can be gamed too easily. I’m personally in favor of only implementing the most robust systems in MaidSafe’s core, the ones that stand the least chance of being cheated. If these popularity mechanisms are being gamed on a large scale it will really harm MaidSafe’s general reputation, dragging down the trust in the other mechanisms with it.
In contrast, the farming mechanism is robust for example. Farming is objectively measurable and if implemented properly, it can’t be faked. You can’t measure popularity like you can measure disk space. Popularity is fuzzy, it doesn’t have an objective and trustworthy unit of measurement.