I've been sitting on the fence a bit trying to figure out what's what.
Having deployed a payment service intended for mass use inside one of the top 3 global retailers, you do need to plan ahead for optimistic outcomes and always keep the end users experience with your product in your sights.
I have come to the conclusion that while you don't need to update now, or possibly not for another year, the sooner you do it the less of an impact it will be for new adopters.
What is being discussed here is that we should wait until Bitcoin becomes more popular before making a hard fork. That's precisely the wrong time to do it.
Do it in anticipation of more transactions so that people coming to Bitcoin won't be reading news stories about a major upgrade that might cause problems.
We also have to consider that nearly $1bn in VC money is going in to development. That investment will soon see new services pouring out and those firms will all have promotional money. The last thing they would want is to start a promotional push, and then face the same situation of having to tell their users that a major update is needed and the update carries some risks.
I think that a hard fork should be made in line with BIP100. The hard fork should carry 1mb blocks as a starting point, but with some dynamic scaling options using just in time delivery to enable soft forking for increasing blocks sizes when capacity increases are required.
This should be part of a series of moves that seeks to take Bitcoin out of Beta. Removing the limit means that the protocol can scale to Visa levels and beyond.
The introduction of Sidechains and Lightning networks means that capacity can be co-managed with micro-payment and off chain channels doing some of the heavy lifting. A dynamic model allows for these systems to work alongside Bitcoin block size planning.
Bitcoin is no longer a proof of concept project. The concept has been proven. Bitcoin Core 0.11 should be the last beta release cycle. Sidechains and Lightning type networks should be dovetailed to work with an eventual Bitcoin Core 1.0 production ready release.