I hate hate hate the whole "30 levels in three months" thing that 3.x and later editions encourage. Adding training time (if it is significant) suddenly makes a huge difference. What if it takes 1 week per level you are going to gain to train? Then by the time you've attained 30th level, you have spent nearly nine years training.
Just a few thoughts...
I agree that I don't like the leveling up in a single year's game time. The way I do is more like the old RPGA style of issuing out time units for each adventure and then stating that it takes X amount of time to complete an adventure, X amount of time for players to do their own thing like profession checks, perform checks, crafting checks, and making magic items. Adding a training time can also be something that the DM adds.
For DM's who want players to train before leveling up, I would recommend to not grant any XP during sessions until the adventure is over or reaching a certain "save point". For example, here's my thinking:
At 1st level, the PC's are hired to explore a ruin and recover a magic dagger. Unfortunately, the ruin is swarming with orcs.
From the outset, the DM determines that the PC's will spend a month on this adventure. So if the campaign starts Jan 1st in the year 1300, then once the adventure is finished, it's Feb 1st, 1300.
Secondly, the DM factors in a time period for player characters to do other things aside from adventuring. Got professions, performance, craft skills? You get to use them in this time. To make it simple, the DM decides that this represents another month. So at the conclusion of this adventure, it's now March 1st, 1300
Thirdly, the DM now tallies up the XP for the adventure and determines that the PC's kill haul is 1,500 XP, enough to make 2nd level. Now the DM determines that time must be spent making arrangements to level up. He determines that this is another month. Now it's April 1st, 1300 and the PC's have played one adventure and three months have gone by. If the campaign goes from 1st to 30th level, it will take roughly 87 months to reach 30th level which comes to just over 7 years. Now if we have the PC's sit out the winter for three months out of the year, I believe that will almost three years so a campaign will run almost 10 years from 1st to 30th level. Of course, if we go with Jester's excellent 1 week per level suggestion then we just added created 9 years for training alone and now our campaign goes from 10 years to almost 16 years campaign.
So what do we do about mega-adventures and time-sensitive adventures--or worse mega-adventures that are time-sensitive?
Well, let's break it down by the parts first.
For the mega-adventure in which players are expected to gain several levels (at least two), the DM must identify a "save point" in which to do all the stuff as above. If the PC's must journey to another town halfway through the module, then that's your save point in which the DM awards XP, states X amount of time has passed and begin training. It's important for the DM to make notes of when that point will occur.
For the time-sensitive adventure, then handwaiving that taking X amount of time for each adventure, isn't going to work. The DM will just need to keep track of the actual time expended toward the conclusion of the adventure; however, at the end of the adventure, he can backload extra time for players to handle their own personal activities and training so that it still equals out a three month period of time or whatever the DM sets.
Now for the mega-adventure that's time-sensitive, well, this is probably going to be the big challenge for the DM. This probably an issue where there is going to have to be some kind of tradeoff. If the DM is really hardup about instituting training time, then he should factor that in to allow for the end date to occur later or allow players to waive it while on this kind of advenutre. Personally, I avoid creating the mega-adventures that are time-sensitive and opt for ones that are shorter and don't reward much XP.
Happy gaming!