I think one of the main reasons people choose to use jtoomimnet is the excessively increased new txn size/share limit which allows producing larger blocks at the expense of DoS vulnerability.
It's not excessive when the new transaction size per share limit is pegged to the Bitcoin network's block size limit. And it should be pegged to the Bitcoin network's limit.
There have been many instances of Bitcoin blocks being found within seconds of each other. One such example is
this recent block by ckpool.org that was found a mere 19 seconds after the previous block, which incidentally is well below P2Pool's share interval of 30 seconds. If a mainnet P2Pool miner were to have found ckpool.org's block, that block would have had only a little more than 100 kB of transactions in it, because there was only time for one P2Pool share to be found, and so only
100 kB of new transactions plus whatever transactions can be found in the sharechain that weren't already used by Bitcoin after the most recent block, within that share. As it is, ckpool.org, without such an arbitrary limit, mined a block that was 918.59 kB in size and 3,674.12 kWU in weight, even after only 19 seconds from the previous block.
To put it simply, with mainnet P2Pool's limit of 100 kB of new transactions per share and with each share taking approximately 30 seconds to find, it would have to take mainnet P2Pool around five minutes to mine a full 1 MB block.
Not only are you shooting yourselves in both feet with this artificial limit (in terms of deliberately limiting how much transaction fees you would earn from blocks found in less than five minutes), but you are also intentionally placing a bottleneck in Bitcoin's transaction-confirmation process (and doing the Bitcoin network a great disservice) by intentionally not mining full blocks when
you could have.
(Do not confuse
full with
big here, lest you accuse me of being a so-called "big-blocker." I support, and intend to continue to support, the Bitcoin Core chain.)
Edit: The third sentence in the second paragraph has been amended from "and so only 100 kB of new transactions and whatever that was left from mainnet P2Pool's 2.5 MB mempool after the previous block was found, within that share" to "and so only 100 kB of new transactions plus whatever transactions can be found in the sharechain that weren't already used by Bitcoin after the most recent block, within that share" to more accurately explain P2Pool's current working design.