I think some of the "heat" regarding the block size limit and multiple-implementations comes down to the existence of two opposing visions of what Bitcoin is.  Some think Bitcoin is ultimately governed by mathematics, others think it's the market.  

My opinion is that Bitcoin is ultimately governed by the market; only over the short term is it governed by mathematics (the source code).  Because this is my view point, multiple implementations are a good thing to me because they remove friction and allow the market to more easily meet its demands (e.g., a larger block size limit).  If someone has the opposing view, then I could understand why they might view multiple implementations as a danger (e.g., the rules could be changed by the market).

What's interesting is that the answer to who's vision is correct is, at least partly, testable.  If the market-governance theory is correct, then it should not be possible to drive a fee market significantly above the free-market equilibrium.  The market will tolerate a block size limit to the right of Q* (i.e., a limit serving as an anti-spam measure but not affecting the free market dynamics).  



However, the market will not tolerate a limit to the left of Q* (a limit that results in a deadweight loss of economic activity).  



If the market wants to be at Q*, how can that same market force it to be at Qmax instead?  

If the market theory is correct, then the total miner fees should never significantly exceed the total coins lost to orphaning over a sustained period of time.  This would be one way to refute the market-governance hypothesis.  Since Q* is now very close to bumping into Qmax, perhaps we'll have our answer within a year or so.


Cross posted: https://bitco.in/forum/threads/block-space-as-a-commodity-a-transaction-fee-market-exists-without-a-block-size-limit.58/page-6#post-1944