An article I read said that if all credit card transactions in the united states were done in Bitcoins, the block chain would be 600GB per year. If that is the case, then Bitcoin will not be used for day-to-day transactions. I think in ten years I will be 3D printing my food...well maybe more like thirty years from now.
The block chain is probably going to need to be in RAM due to access issues, but 600GB per year is easily within the grasp of what a modest sized outfit can do today. Probably within what many private class users could achieve in the not to distant future.
My main beef with the soothing 'official' scalability projections are that the sights are probably to low in just meeting the credit card use rates. For every credit care transaction I do, I probably do 50 cash ones, and I don't think that credit care use is as high in most parts of the world as in the US. (The other thing which bothers me is that already the system is displaying instability at the current tiny use rates and development speed is probably not nearly high enough...begging support from well capitalized entities to lend a hand...)
So, it seems pretty pointless to try to make native Bitcoin be a dominant transaction currency solution. That leave two choices:
1) Keep it as a tight 'reserve currency' that can be secured by a large number of autonomous users operating on disposable budgets throughout the world, and ride 'exchange currency' solutions on top of it, or
2) Grow it just enough to exclude everyone with lower capabilities they yourself...then build second tier exchange currency solutions anyway.
I personally wish to see #1, but as they say, 'if wishes were horses then beggars would ride.'
I project that what will happen will be the gigantic corporations and governments will succeed in growing the solution to a size which will exclude other players, then operate it for 'free' to users while extracting the intelligence info which is worth a lot more to them.