Quick poll. I'm implementing the time expressions for Muse (you input a time to the history generator). There are two possible options.
The first option is UnixTime. UnixTime is my real time option - where your world continues to evolve while the game isn't running. UnixTime returns the number of seconds since midnight, January 1, 1970. It constantly, reliably, moves up by one. Moving by one is good for this purpose, because that means the changes will subtly change over time rather than sudden radical changes. The problem with Unix Time is that due to memory constraints, in 2038, it will fail. I don't know why this bothers me so much, as by then Construct 5 will be pumping out much cooler games than this, but it bothers me enough that I'm implementing my own special version of UnixTime I'm calling MuseTime. In UnixTime, midnight - January 1, 1970 is referred to as the "epoch". MuseTime will be referred to as an expression, much like UnixTime, with one exception. You will be able to set your own epoch. That way, you can set it to whenever the game first begins, or whatever, and the player has a good solid 68 years before it loops back on itself. (Which would cause a sudden change of history within your game.) If players are playing your game for 68 years - I doubt you'll be worrying about this.
The second option is counting the number of seconds your game has been running, in which case the game will pause while you are not playing. This will come with "Punch In" and "Punch Out" actions to start and stop the clock. My question to you guys is that this is going to need to save data. Would you rather this data be serialized with Construct's native savegames, controlled via the "Save" action, or would you rather this data be saved in an external file. Personally I've never used Construct's save action and have no idea how and if it works, but, and I've always wanted to say this: modifying that external file WILL result in unexpected time travel. That's something you probably want to avoid.
Anyway what do you think?