Rage

It's also an id game, so it will be unpretentious and direct. "Matt Hooper - he's kind of my partner in crime on this - and I hate playing videogames where you get an objective and you have no idea what to do," says Willits. "We want to know where we need to go, who we need to talk to, what we need to find and bring back.

"So yes, you will have a classic videogame interface where you know where you're going, where you have a map with a dot on it. People want the potential to explore and do things at their own pace, but they also want to know what's going on." So if we're playing Rage and we're given objectives, we can go off and explore, and later rely on the interface to bring us back to task. "Absolutely."

"We really want Rage to be more approachable and accessible than even Doom 3 was. We want to ramp it up for the hardcore but keep it accessible to the novice," Willits elaborates. Carmack will add to this later: "It's going to be balanced so that we never want to frustrate the player, you never want to make them do something where they're upset, pissed off at the game, whatever, which fundamentally means that you can't challenge them too much because there's a big trade-off there."

Not that they're illustrating that with too many examples. Asked for details on specific missions, Willits demurs. "We don't want to dig too far into the game because we have quite a while before it comes out," he says. But he goes a bit further. "We're not going to be forced and limited into this linear, same thing over and over again structure. We know that players are very ADD these days!"

There's almost no discussion of the control layer, either (except to say "360 is our primary target" in terms of input), but with id games you can probably take some of it for granted. It will feel like a great FPS game: a mixture of player-guided precision and fantastical possibilities; responsive and probably inhumanly fast to manoeuvre through. Willits confirms "the game will be tweaked and adjusted for the keyboard and mouse", which, as Carmack said last week, remains the best option for FPS. PS3 offers keyboard and mouse support, of course, but Willits says he's not sure whether that will make it into Rage.

Elsewhere, oddly for id, there's more discussion of what won't be happening with Rage multiplayer than what will. The co-operative elements outlined last year remain in place: a separate, fun-first two-player co-operative mode (with split-screen options on console), which unlocks rewards that feed into the single-player.

But competitive multiplayer is up in the air. "We haven't quite nailed down exactly what we want to deal with that aspect of Rage," says Willits. "John has some ideas, I have some ideas, but to be honest we haven't really worked through that very much." For Carmack's part, he says Rage won't take the structured, persistent, stat and leaderboard-heavy approach of games like Call of Duty 4 and Halo 3.

But to come back to what we do know, we do know about the technology. The levels of detail thrown about by the id Tech 5 engine on PS3, 360 and PC will be equivalent - and locked at 60 frames-per-second, everyone insists. The engine does something called virtualising textures, which through Carmack-architected voodoo allows the artists to disregard traditional texture-memory boundaries and keep piling on detail throughout - as evidently they have done.

Then again, id's fans expect the games to look great and handle comfortably. It's one of the reasons we're talking to Willits and Carmack about Rage a couple of rooms away from 5,000 adoring fans. What makes Rage interesting - and worth flying into Dallas' 40-degree heat to check up on - is that a lot of it is outside id's comfort zone. A relatively open world, driving elements, RPG-style progression, questing. After two years of trailers and extensive interrogation, we can only guess whether or not it will work out in the end.

What we do know, though, is that it certainly wasn't a boring option. "When it's fun and when it's done," says Willits in answer to the obvious closing question.