A disclaimer:
I'm not fully up to speed with the Forge theoretical framework yet, and I may omit or misuse the appropriate terms of art. I hope to learn by doing & appreciate corrections and instruction!

Some relevant history:
I started playing D&D in 1980, when I was 10, and had more-or-less stopped by 1985 (and near the end of this time Call of Cthulu had been the system we played with the most interest). Only one of my gaming friendships from this time survived high school, and he and I would occasionally get together some one-on-one play, the most memorable of which were set in Leiber's Lankhmar.

I set out to get involved in RPGs again in 1993 by answering ads for games in Chicago area stores. Many of the ones that I checked out were D&D, but the group that I joined was playing Ars Magica. Part of the attraction was System-Does-Matter; the way that developing a character's investment in the various magical nouns and verbs has a direct effect on the sort of personality that character will develop through gameplay alone impressed me greatly.

The other thing I really liked about Ars is that its fantasy elements were played off against a hard historical background: as a fan of Tim Powers' novels using this technique, I enjoyed the way our saga created the "secret history" of the 990's. More than that, though, I liked the way that everything you could learn from any external source was valid material for the game, so that every player could have the pleasure of digging into research and using it to enrich one's play. The more that a game's background is "soft" or subjective, the more this pleasure is restricted to the GM.

What I'm finding in the D&D 3.5E campaign I'm currently DMing is that for players of my loosely-defined generation, the D&D default setting (both common assumptions about what D&D is and the encoding of the setting in the rules) serves as a "hard" background. About half of my original group had played AD&D in junior high and high school, while others (including all the female players) knew people who had played but never did so back then. My comments are most largely true for the former group, but even the latter have had the D&D archetypes seep pretty deeply into their consciousness through the cultural groundwater.

For the basic reason that everyone knows more or less what it is, I've found D&D to be excellent for (re)introducing people to roleplaying, and have been surprised by how often I meet people outside of gamer contexts who remember fondly their past history of D&D.

More than that, however, our actual play involves fruitful rebounds off this background of what-everyone-knows. The uninteresting examples first:

#1. Metagame Discussion Squelched by Social Contract.
Some new players join the campaign who are more experienced roleplayers than the original group. One of our habitual metagamers asks "What level are you guys?" when the new players introduce their characters. They respond in-character "I don't know what you mean" : an appropriate definition of their preferred play style, but nothing new.

#2. Metagaming as Nostalgia.
As I was re-capping the backstory of the campaign & describing the party's explorations beneath Zenopus's Tower, I held up my battered childhood copy of Holmes' blue book basic booklet, making the point that our game had grown from the seed of that original sample dungeon. The new players were suitably impressed, but this didn't affect actual play.

Now the interesting ones:

#3. Metagaming as Player Tool for Enforcing the Hard Background.
"What level are you" doesn't have an in-game answer and so is best brushed off in character, as per example #1. When players were first asking NPCs "Are you good or evil," my knee-jerk response was to do the same. I'm glad I didn't, because it has been very stimulating for me to have to take seriously the fact that these terms do have defined game meanings known to characters and players alike.

My first response was to create a temple which offered a certification program: people in positions of responsibility are expected to have a yearly inspection on file that they are Good. It then developed that the pirates from the sample dungeon were running a scheme where they erased the most recent parts of people's souls, allowing them to pass the temple's certification by removing the stain of sin (but not the inclination to sin again, so the pirates did a lot of repeat business).

This good-and-evil dialogue has been much more fruitful because it is implemented in game terms that bind me as the DM as well as the players: rather than just BSing about the nature of evil, they can cast spells, experiment with the pirate soul-eraser, and otherwise use the gameplay to explore the objective meta-framework of alignment.

#4. Metagaming as Foreshadowing & Playing Against Expectations.
One of my players is the kind of guy who read every module there ever was back in the day. I've greatly enjoyed playing on this to set up expectations, such as dropping the name of Markessa from the Slavelords series. Our in-character code for sharing his meta-game knowledge is "An old elven ballad says..."

I like the fact that these iconic characters also exist outside myself & that I can invoke their power at will; I like being bound to the meta-framework of agreeing that anything players know about them from their own research is true. I like having an audience that appreciates the inventions I'm able to introduce within this framework. This goes a little beyond the Ars Magica framework because it has narrative elements as well; it's more like playing a Call of Cthulu game that's explicitly set in the location and timeframe of HPL's stories.

Thanks for listening!
- Tavis

Interesting. On reflection, it occurs to me that D&D, more than any other RPG I can think of, has a firmly established context for its play, and this context is a part of culture at large. Which helps partially to explain (to me, anyway) the game's continued popularity.