Don't get me wrong. I'm not exactly in favor of welfare epics. I actually prefer ways to get better geared players back into lower tier instances, such as Rift's weekly raid quests that encourage even the best end-game players to "pug" (pick-up group) the earlier raids, giving less intense players a chance to see that content and get some gear. Again, this points to good ways (motivations) versus bad ways (gear handouts) to solve a design segregation barrier.
But these systems don't always work. Players are locked to only one raid instance a week to prevent them from getting too much loot, but this means they can't go back and help friends complete the instance, even if they were willing to be locked out of getting more loot. Raid lockouts are another design barrier that might need some rethinking if players are to be encouraged to play together. This is especially important if your goal is to get people who aren't interested in a raid's loot into the raid in the first place. Why offer a special incentive to get high-end raiders into low-end raids, then lock them out of the zone after one run?
Likewise, the automatic dungeon finders that group people and teleport them into a dungeon, featured in games like WoW and Rift, were supposed to overcome social barriers to actually forming groups with players. But the effect systems like these have on building communities is questionable, at best.
Sometimes, the differentiation between design and operational barriers blur, such as in Turbine's Dungeons & Dragons Online. When DDO went free-to-play, it locked many of the game's adventure areas as premium pay-only content, an operational barrier.
But Turbine also allowed paying users who had bought the adventure area, or who were paying for a subscription, to buy guest passes, allowing users who did not buy that adventure to accompany that player inside.
This overcame an operational barrier, but it left a design barrier in its place: what motivation was there for a player to hand out guest passes to strangers? What motivation was there to do that dungeon at all, rather than a free one?
Dungeons & Dragons Online
DDO's guest passes are an example of letting players play together, but what about the getting them to play together part? The "let" is a necessary first step; the "get" is the key to customer satisfaction, conversion, retention, and all that monetary goodness. But here are a few bits of advice to get everyone started:
Operational barriers are generally a Bad Thing. No one likes to be shut out from their friends because of some arbitrary operational setting a producer in a dark, dank cell decided on. Reducing operational barriers as much as possible should be a primary focus.
The free-to-play "revolution" sort of embraces this idea, by tearing down the most obvious of operational barriers: the subscription fee. But the subscription fee doesn't have to go, so long as the product you sell provides good value to all its customers -- including those who opt out of subscribing for a period -- and avoids other operational barriers.
Design barriers are generally a Good Thing. They let players feel a sense of accomplishment and are often the main goals players work towards overcoming in games. The problem is when they get in the way of players playing together.
There are two general ways of overcoming these barriers: the leg-up approach of letting players bypass the barriers, which is generally an unpopular approach, especially among those players who worked hard for the satisfaction of overcoming them legitimately; and the bribery approach of paying off players to go back down to that stuff they had done before, which is generally popular and, if well implemented, encourages building those lasting social community bonds that are so integral to building a strong game in the first place.