The scale of changes can be an issue. I used DDI extensively with 4E, but if I'd gone to a game with my Player's Handbook, a fair deal of it was obsolete when I got there. (This is especially true of organised play, which uses the most up-to-date rules). While I really appreciated the attention to balancing the game, it did progressively take away a lot of the value of the PHB.
This is really only part of my concern. I really don't mind the three year cycle (say Savage Worlds or D&D 3.0 to 3.5). It might be a little close together, especially if the "core three" release spans six months every time or there are bigger changes or lots of supplements I'd want updated. Five to seven years might be better, but it's somewhat subjective. Five would probably cover the cultural ADHD, without feeling like we're on a treadmill.
What I mind is not having a clear line of demarcation between when rules are valid and not. As a player, I don't want to be in a position where I spend time making up a character only to find out that the class has been modified so that I have to redo some of the work. Worse still, I cast a spell only to be told that it's been "fixed" or plan combat tactics based on things like attacks of opportunity that have been tweaked. Those are the sort of thing that would get me to walk out of a game. As a DM, I don't want to put my players in that position.
The flip answer is "Don't use rules without discussing them with the group." That's all well and good, except for the new player. Or if someone forgets to make a note about something because it doesn't impact them, at the time. So, the next flip answer is "Don't change rules during a campaign," which is functionally the same as designating an edition break.
Here's a model I could get behind, though: Resurrect Dragon magazine -- preferably as print-or-PDF, like my MSDN subscription is. Make it something I can save for posterity. This is an index-ready reference for the cool ideas. Hopefully, they aren't just flailing about, and we'll actually get some reasonably complete options out of the deal. Then, the DM has a list of options, not something that can change day-to-day. Once there are enough changes and they see wide enough adoption (or whatever measure of "this is better" consensus), then update the edition.
If there are some groups that are all online enough and want to participate in the "beta" rules, then put those up somewhere on the site. When the rule is "done", publish it in Dragon. If they don't feel right about charging for the content, then just turn it into a "bundle" of the web content that I can easily download and save whenever editions inevitably do increment. Really, though, I'm assuming there's some other content that could be put into Dragon to make it worth paying for. That content may be fluff (I still love and reference the old "[race] Point of View" articles from the early 1980s, reprinted in "Best of, Vol 3"), add on modules (NPC classes, feats, magic items), or minor options (realistic height & weight from #91, pain from #118, colorful crits) that are never expected to make it into canon.
Now,
that is the absolute best way I can think of to support this edition. Skip the splatbooks. Publish a few settings, some nice adventure paths, and -- maybe -- a couple of semi-core expansions (psionics fits the bill, IMO, as do the occasional players' companions to the adventures, YMMV). Everything else goes in Dragon.
