I also do not think this will take as long as it has taken the XServer to evolve to its present form for Wayland to either evolve to the same degree of complexity/features/support. However realistically speaking, I don't see Wayland taking over the XServer in less than 5 years time-frame, could take even longer. However, recent tilts into the development of FLOSS software has been blazing fast. Take Compiz for instance, it has taken distros to enable it by default (Ubuntu 8.10 and others) over two years, and its been under development for at least four years (that I know of, anyway), so from inception to deployment of this rather "useless" piece of desktop bling it has been about 4-5 years. I see a similar path for Wayland. It could be faster due to the interest and momentum generated by these other projects.
One thing that has sucked the air *out* of compiz (as a project) is that desktop environments (especially KDE, but equally true of GNOME, let alone the lighter desktop environments) are including their own compositing managers which are doing their darndest to obviate (either partially or completely) compiz. (Look *just* at KWin's rather lengthy, and getting lengthier, list of natively-supported desktop effects as long as you have OpenGL/XRender/DRI support - how many of these are duplicates of those normally brought to the table by compiz?) Desktop environments (whether heavy or lightweight) are becoming more and more monolithic - in fact, they are becoming a lot like - dare I even say it - *Windows*.