Coz he is more worried about the problems with p2pool and doing something about them, rather than sticking his head in the sand.
Increasing the new txs size/share limit is sticking your head in the sand about attack vectors. Unless of course your use case is a private pool aided by other hashers.
Well if you consider than 39% of blocks will be less than 1MB, then p2pool can't compete with any other pool that always mines full blocks ... and that's of course ignoring the other problems he's trying to resolve.
Your basically scamming people when you say that they'll earn more on p2pool, since they wont for a number of specific reasons including:
1) mining fee of between 0% and 1% (varies per block found and if you run your own node or use someone else's)
2) lost transaction fees since 39% of blocks will average under 1MB, though that's no longer the block size limit on other pools
3) inequalities for all but the largest miner on p2pool - unlike all private pools that don't have their miners battle over submitted shares, no matter the size of the miner
4) current pool size gives an expected share payout based on multiple diff changes - which are 'usually' increasing, though at the moment there appears to be a short term of it dropping
So basically your saying p2pool has an attack vector that implies no one should use p2pool - since either
1) jtoomim's changes that resolve some of the above are dangerous
2) your version ensures poor payouts due to all the above
Well that's an unexpected thing for you to imply about p2pool ... ... ...