@Inca you forgot point 4).
They are going to remove the space reserved for 0 fee txs (reserved for "old" coins). It used to be there since the beginning as already noted on this very thread. Software changes that implement that are already merged.
The main motivation for such change is said to be reduction in code complexity. They sure are funny.
They are outpacing any sane engineer-rules-of-thumb timetable to implement SegWit and deploy it via a soft-fork.
A change that will give us a mere 75% more block capacity at the cost of changing the system core data structure, the Blockchain, without even an official BIP, but at the same time for the sake code simplicity they are going to ditch a feature has been here form day one?
Isn't this a clear example of Doublethink?
You know the funny thing about it? Luke-jr want to revert it, maybe due to the fact it has a lot of "old" coins that could be moved without fees under the old rule:
https://github.com/bitcoin/bitcoin/pull/7151
and he's encountering quite a bit of resistance. It will be funny to see how this will play out.
That said, I think there would be significant demand for BlockStream's payment layer solution even if BIP101 was deployed. Instant-visa-scale payments' txs will always fit better on a LN rather than on the main chain.
Sure such LN market will be an order of magnitude (maybe two) higher once full RBF will move from opt-in to opt-out and SegWit ( i.e. 1.75 block size limit and fixed malleability).
I really do hope that BU / Btcd / Investors backed full nodes will gain enough market share to remove such concentration of power from Bitcoin Core devs. The sooner the better.