I also view them as doing an VALUABLE SERVICE for Bitcoin and don't think it should be disrupted.
There is definitely a valuable aspect to SatoshiDICE but it also comes with a cost. If they simply stop sending spam transactions to inform players of a loss, the problem would go away. From what I understand, the attitude from the principals was "if you can't deal with it then Bitcoin doesn't deserve to exist."
Without Satoshi Dice, we are left to guesstimate what will happen when we start reaching limits.
That's not true. First of all we have the soft limit, which we are consistently hitting. Second, we already have a pretty good understanding of what will happen when we reach limits. The fees will go up. Which is exactly what is happening now. Ordinary users have to pay higher fees just because SD is spewing 80% of all tx on the entire Bitcoin network. 10 years from now that's no big deal, since blocks will be full all the time.
But to have the higher fees now, much sooner than needed, can only choke adoption not spur it. And we need that adoption rate now a hell of a lot more than we need a gambling site that consumes many times more resources than normal.
With Satoshi Dice, we reach those limits in a non-committal way.
If by non-committal you mean "forced to store every losing transaction notification with unspendable and unprunable outputs from now until the end of time" then yeah I guess...I prefer the dictionary definition of non-committal.
We get to find out how Bitcoin reacts under load, using a load that is for all intents and purposes optional.
An artificial load of transactions with unspendable and unprunable outputs that compete with legitimate transactions and drive fees up without a corresponding increase in the size of the user base. Judging from the rate of some of the transactions and the wallet they came from, it is very likely that bots are doing much of the betting.
One could argue that implementing the patch lets us "get to find out how the development community reacts to economic attacks on Bitcoin's finite resources."
we always have the option of throwing out the Satoshi Dice noise long enough to re-engineer Bitcoin to handle more activity.
This is not an activity problem. Once Bitcoin reaches a critical mass (when the blocks are always full) this problem will solve itself and we wouldn't need the "ban" or any sort of client fix. It's only a problem NOW in the early adoption phase because it drives up fees for no good reason.
Eventually fees will go up no matter what, because blocks will be full and transaction space will always be a limited resource (even if we increase the limit). But having fees go up now only hurts the adoption rate, which we need more than we need to support the handling of every gambling addicts manual or automated pull of the slot machine arm or roulette wheel.