It took them 5-6 years to makes the current engine/game. So even if they started making the new engine and 2.0 the week YoshiP took over in dec 2010 that would mean they are shooting for a new engine and game in about 2 years.
5/6 years = crappy game + crappy engine
2 years = Awsome game + awsome engine?
Added to this is the fact that FFXIV will have roughly 20-50k players with 10+ 50s most of which will have 10s-100s of millions of gil. If you are expecting a happy ending I would say you are a very very optimistic person.
Well, I will most likely have 10+ level 50 jobs, can't argue with that, but 10 million gil? Probably not... though that's kind of a nice goal, I might have to try for it now:).
The one thing to understand is the use of the word Engine. In the world of software development words like server, service, engine, and application are fairly ambiguous.
You have a physical server, which runs a web server, which has web services that report the results of different services... Each item being a different layer of disassociation through relative use.
Engine is very much the same thing. An average game engine is made up of a physics engine, a graphics engine, a combat engine, a HCI engine (Human-Computer-Interaction, aka your key bindings, controller input mapping, etc) and data management engine (ok this last item is more specific to Multi-user/persistant data systems).
Now, the overall Final Fantasy Game engine is actually fairly impressive, and giant steps forward from 11. However, there are very glaring key elements that are so poorly designed/implemented that they detract so much from the game play they can make the whole game seem to be a giant pile of don't touch.
However, from an unbiased look at it, you have the key bindings. You have the response rate. The render requirements. Object visibility. Latency prediction. Data retrieval. Movement. Abilities (static, non-unique, over-designed). Mob reactions, AI Logic. Region Scaling. Collision Detection (every game could always use improvements with this). Instance handling, Quests (or lack there of). Linkshell interaction, Chat usage, Party mechanics, crafting interface, gathering interface, NPC interactions. (I am sure I missed a couple, but it's a pretty good list of currently identified flaws)
Now, considering how each one of those elements falls into a category of identifying the flaws, design desired interaction, code desired fix.
Considering each item mentioned had to be initially scoped, designed, implemented, debugged, and QCed before the original release (5 years of development), including the other portions of the game (Mapping, character, Mob, NPC Art, among a list of enormous quantity), two years compared to 5 is not that unreasonable. Considering they are trying to re-use as much as they possibly can, and improve on it. That which can't be improved is replaced (see Market Wards).