Thank you all for a very insightful discussion. I would answer earlier but was travelling to Saint Petersburg – to the first Russian Wolfram Technology Conference. I must tell you it was great, Stephen gave an exciting keynote and was besieged by an ocean of questions – I estimate for about 2 hours. But that’s another story. 

Steve Christensen, is indeed a moderator here, along with Bruce Miller and myself. MathGroup will still function for some time until we naturally absorb its audience into the Wolfram Community. Its content will be always available on the web.

What concerns user-founded communities, I think they are great, we value them very much and we can all coexist very harmoniously. For example I am quite an avid member of Mathematica Stackexchange, was there at the founding times (13th person to register for area 51), and so are a few colleagues of mine from Wolfram. We enjoy that lively place. All communities have their natural focus and demographics. And this is why they are quite complementary to each other. Wolfram Community, for instance, provides space for discussion about Wolfram Technologies beyond Mathematica, such as System Modeler or Wolfram Alpha, or beyond Q&A format, which is not acceptable at Stackexchange due to their rules. I hope this will help Stackexchange to lower the rate of closed “off topic” questions by redirecting them to the Wolfram Community. We also encourage real-names and real-bios here to put discussions in a professional occupational context. And of course we envision our web-enabled technologies, such as Mathematica Online or CDF, to be integrated into the Wolfram Community and vice versa in the future.

By the way, I think there are/were actually many more communities than we (or at least I) have expected – inside of larger social infrastructures. For example, LinkedIn Mathematica Users group (2,857 members), Reddit Mathematica (> 1,850 members), Quora Wolfram|Alpha (> 5,800 members), and I suspect even more. With recent boom of personal blogging we should consider that as a form of community too, - there are quite a few Mathematica blogs out there (e.g. matthen, archery). All forming their natural niche for a specific demographic – which is simply put great.

What Craig said at the end seems a likely direction. I can give some tentative ideas, that should be taken only as that: none of this is definite and none of it is in my control. So, while I have a small amount of insider knowledge, I am not speaking for Wolfram Research.

(1) The Community will not in any way merge with Mathematica.StackExchange.com. It's not a possibility (they are independent of Wolfram Research). Even if it was possible it would not be a good idea for any of several reasons.

(2) It is certainly possible that the Community might subsume MathGroup. One motivation for this new forum was that it would retain similar Q&A features as the group, and a similar style of moderation, but would reflect the capabilities of the newer sites that have appeared in recent years (e.g. code sharing, which is in the works, upvoting, and more).

(3) (I hope this is not confidential in any way...) To the best of my knowledge, Steve Christensen is now also a moderator for the Community forum. If we do go the route of merging with MathGroup, I believe it will be a relatively seamless transition and one that has Steve serving in a similar capacity. Though perhaps (I should hope) dealing with less spam than at present.

(4) I'll be up front here and note that (3) above is not the complete story in one important detail. MathGroup, as run by Steve Christensen, has had considerable autonomy as an entity largely independent of Wolfram Research (though it is more closely connected than the StackExchange site). Clearly the Community forum is under the direct auspices of the company. This distinction has posed an issue at least in theory. In practice I think the main idea for both MathGroup and Community is to apply common sense to moderation. So the question of autonomy might not be much of an issue in practice. That at least is my hope.

(5) MathGroup also exists as the Usenet group comp.soft-sys.math.mathematica. On the one hand, I do not know whether or how such a group might be "subsumable" into a very different type of existence. On the other, my impression is that many Usenet groups are losing the popularity they had through the 90's and early 00's. While MathGroup remains fairly strong it is not obvious that it can continue in its current form indefinitely, as more modern forums are simply replacing the Usenet model. I guess my point here is that, independent of Wolfram Research vs independent control, there are external issues that both make merging tricky, and that argue for doing so in the not too distant future.

Vitaliy Kaurov knows more about the plans for the Community than I do. He is traveling right now, which is why I am attempting, in a perhaps clumsy manner, to address questions and concerns raised here. I'm hoping he'll post corrections within a few days if I made any major mistakes.