In Lusternia, an online text game I frequently play, there are examples of non-communication common in any online community (Use of capitals, pauses, or emoticons), but many of the other uses are a form of metacommunication.
In the game, characters communicate actions to each other to simulate real-world non-verbal communication. For example, descriptions of facial expressions, posture, pitch or tone of voice, eye contact, etc. All of these are deliberately chosen by the players to convey a particular mood or emphasize a ‘spoken’ thought- in other words, players are required to interpret non-verbal communication in everyday life and decode it into written language. The receiver(other player) must then correctly interpret the words into the meaning associated with the non-verbal act (I.E., smiling expresses happiness). Individuals skilled at reading other’s visual cues in public settings- and writing- tend to be better at this.
Because the communication is direct, expressing a particular non-verbal cue is often faster, but there is difficulty trying to replicate the degree of information that comes from a face-to-face conversation. Attempting to describe tone, posture, gestures, eye movements, accents, expression, etc., at the same rate as we would normally receive it would result in too much information, and be impossible to keep track of. At the same time, the lack of those cues can result in miscommunication and confusion.
Negotiating that balance requires metacommunication about what is considered the most important in non-verbal communication. To make it easier, standard conventions are created within the game: Certain ways of expressing actions and emotion become preferred(called pre-set emotes), and an understanding about the flow of conversation/actions replaces turn-taking signals, or non-verbal cues as to whose turn it is to speak. Generally, one player speaks or acts (a short written paragraph), and the other waits, alternating back and forth.