And what's the alleged advantage? Faster transactions?
Well, yes, but not as secure as 10 minutes blocks. With the same computing power, a 10min block will be twice as secure as a 5min block. You're not more protected against double spending with 2 5mins blocks than with one 10 min block.
Also, we don't want faster transactions, we want real time transactions, and that cannot happen within the block chain. Other means are needed, for example, the green addresses technique.
I don't think 5min blocks are a good idea.
I have so much of the network hashing power that on a ten-minute block system, say Ten-coin, I have a 1% chance of producing a block in one minute. Therefore the amount of produced blocks in ten minutes on average is 0.1, and the rest of the network produces 0.9.
I switch over to a five-minute block system with equal hashing power. I now have approx. 2% chance of producing a block in one minute. Therefore the amount of produced blocks in five minutes on average is 0.1, and the rest of the network produces 0.9.
Since double-spends require more than one block in a row, and you have the same share of blocks in a five-minute system, the chance of more than one block in a row is the same or marginally increased.