It seems like I might be the only one who has serious mixed feelings about this. For one thing, I think some posters didn't read the article carefully enough.
If we do identify a problem area, the next step depends on the magnitude of the change. Some alterations are simple and easy. If a rule is unclear, we can include an update in a FAQ or similar resource. If a number is wrong or a rule is missing a keyword, we can update future printings of the relevant books and compile a list of errata. For changes of this magnitude, we’ll aim to provide annual updates. We’ll make actual rules changes (as opposed to updating a FAQ) only when absolutely necessary. If players and DMs feel they need to replace their books because of these changes, we’ve gone too far.
Many problems are not as simple to change, however.
In other words, Mearls is distinguishing between two different kinds of updates: the "simple and easy" kind, and the "greater magnitude" kind. He says that the "simple and easy" kind shouldn't make you feel like you need to buy new books. That line does not apply to the "greater magnitude" kind of changes. Here is what he says about the "greater magnitude" changes:
A revision significant enough to require serious changes to printed books should offer multiple obvious improvements to the game. If you’re buying new books, it should be because you want to—not because we’re twisting your arm. In an ideal world, updates to our printed products should simply capture the incremental updates and revisions that have proven widely popular.
He says he doesn't want anyone to feel strong-armed into buying new books. But an entirely new version of a core class—which has prove widely popular and is now integrated into a new printing of the PHB—is
exactly the thing that would make me feel strong-armed into buying a new PHB.
That isn't what I want. I want the designers to be confident enough in the rules that they're publishing right now that they won't ever replace classes or other major game elements wholesale. Obviously there will be quirks and problems, but I was hoping that the whole point of the public playtest was to ensure that there won't be any quirk and problems serious enough to warrant completely re-writing a class, and publishing the new version in a revision of the PHB. Mearls can claim all he wants that what he's describing here is significantly different from the 3.0 > 3.5 transition, but it isn't. I'll feel obligated to buy new books just the same.
I would much rather have their assurance that errata will be kept to an absolute minimum. I would much rather have a promise that all of the errata for the PHB ever published will fit neatly onto a single attractive page that I can print out and then tape into the back cover of the PHB.
On the other hand, even if a few major overhauls get integrated into the PHB, I can see the benefit of that approach over simply moving on to a completely new edition in 5 years. I would prefer that 5th edition last for 20 years without any "major overhaul"-type changes, but that might be unrealistic. It seems like the choice is really between (a) I have to buy the sixth edition core rulebooks in 5 years because we've moved on to a new edition yet again, and (b) I have to buy new fifth edition core rulebooks in 5 years because a number of "major overhaul" changes have been integrated into them.
Given that choice, I can see the benefit of the latter. I have to re-buy a bunch of expensive books either way, but at least in the latter scenario I'm getting an extremely polished version of a rules-set, whereas in the former scenario I'm just moving on to another new rules-set full of problems.
