In a role-playing game, there are two types of communication taking place: the actual role-playing, and the out-of-game rule discussions. In a tabletop environment these are necessarily blended, but in a digital environment we can now separate them for increased immersion, and this is what we'll do.
The ideal setup would be the following. We are all on a Discord channel but with voice set to push-to-talk. This way, for the majority of the session, the channel will be dead-silent without even any sound from clacking keyboards and so on. You'll be able to eat and drink even, if you want, while playing, without bothering anyone.
Meanwhile, all the role-playing will be happening via text chat. I think Fantasy Grounds provides an interface for that. I will be copy-pasting text descriptions from the books into the chat, and these will blend perfectly with my own descriptions and all the NPC dialogue, and you won't be able to differentiate them the way you would have in a tabletop session, where it would be quite clear when I was speaking of my own accord or reading out a lengthy passage. My delivery of the latter would necessarily seem more artificial than the text appearing suddenly in front of you in the chat, and you reading it to yourself with your inner voice the same way you read a novel, or anything else that comes up in the chat—including your companions' words. The net result will be to make you feel as if you're playing a videogame, but one in which the AI is insanely advanced and all your companions and NPCs behave like normal people, with no repeated lines or simplistic dialogue trees involved. And when at any point you need to ask a rule question, or anything at all that's outside the role-playing, you push the Discord button and speak with your voice to the entire group. When anyone is speaking, the game should be considered paused, and no one should type anything until the discussion is over. All this has the added advantage of removing our "voice-acting" so to speak from the game, even the DM's. It's far less immersive when you hear a single voice playing a thousand roles than reading text on the screen coming from a thousand characters, and even for the players, their accents might not match up with those of their characters, and text communication solves that problem.
Now the above is the ideal implementation of what I am talking about. I have no idea how practical it will be, so we will try it and see. And even if it works alright, we'll still need to get used to it. People will need to learn to type not too little or too much, not to post all at the same time, and so on, which is much more delicate business in text than in voice chat. But I believe we should be able to pull it off, and when we get used to it it should take the game to a whole other level, and really approach that sense that I am aiming for of playing an insanely advanced videogame.
All the above is said for the Fantasy Grounds portion of the game, but I believe I saw somewhere that TaleSpire too is adding text chat, if it doesn't have it already. It is a more complex program though, so we'll see how it works out. I am hopeful in both cases, and I believe with enough patience we should be able to overcome all problems and master this system to attain incredible results.
P.S. Joking will also use the Discord voice chat, unless you're joking in character. This may end up being the most challenging aspect of all, as there's a lot of joking around a D&D table, and I am not sure how well it will translate to Discord as jokes often involve gestures and so on. So I expect this dimension to diminish, in exchange for increased immersion. And we will each judge if the tradeoff is worth it after some time playing the game this way. It's not like we'll have an alternative though, as we won't be able to play in person anyway, so we may as well make the most of the situation and derive maximum benefit from the increased immersion.