I'm mildly alarmed by one trend I'm seeing in the current stretch goals for this game on Kickstarter: You seem to be planning one companion per available class. Now I know you guys at Obsidian know what you're doing, but I still feel as if I should mention this.
You're writing complex, believable characters. The problem with such characters is that, as in real life, it's very individual who you end up liking and who you end up disliking, sometimes strongly. Some of the best fantasy fiction out there has characters that some people love while others hate. Take A Song of Ice and Fire for example. There's only a small handful of characters in that story who the readers can agree about. That's a sign of good writing.
It also represents a problem for games where you have to not only interact with these characters, but also choose whether or not to bring them along with you on your adventures. Let's say there's one class in this game that is best at healing, but the companion that has this class is an unlikeable jerk (which seems to be a common trend with healers for some reason). You're left with only two alternatives: 1) Bring him along anyway and hate every moment of it, or 2) Create a character of that class for yourself, which requires that you know about this character and his flaws before you even start the game.
I feel it's important that we're always given a real choice when it comes to companions, particularly for a game which is likely to take dozens of hours to complete. This means that there should always be more than one companion per character class, preferably two or more.
Dragon Age 2 is a good example of how this can go wrong. There's a very limited selection of companions there, some of which can be missed entirely during a playthrough. When I played the game I never got the kind of party the game assumed I was going to use to have a fighting chance, for the most part because I chose to play a mage. I never had that problem in the first Dragon Age, which showered you in companions.
If Project Eternity was my game, I would have done one of the following:
Eliminate classes and go for a class-less system. This way most companions could be adjusted towards what you most need them for. This is the ideal solution to me for more than one reason.
Add more companions so that the ratio is two companions per class.
Remove classes so that the ratio is two companions per class.
The way things stand right now, the amount of companions as presented on your Kickstarter represents a design issue the way I see it. You may have a master plan that eliminates my concerns, but I'm not seeing it.
Please reconsider the ratio of classes vs companions!