AI Fix, and Budgeting
Hey Folks! Just a short update today, as most of the day was spent on budgeting/planning.
Recently, while testing a build to show a friend, I noticed that some of my AI was skipping over important part of interaction chains. Basically, the AI was remembering that the punchline of the jokes had a big payoff, and wasn't bothering with the setup anymore.
The bug was in the way it decides on what to do next. There's no differentiation between replies and openers in their memory, so they just chose whatever had the best impact in the past. And since that was often the allow/deny response and not the initial request, they just cut to the chase.
So I tweaked the code to only do openers, and the AI seems to be more well-behaved now. The screenshot shows this more natural flow, as opposed to Chinelo, say, accepting cigarettes endlessly from no one :)
The rest of the day (and by "rest," we're talking 7 out of 8 hours) was spent looking at spreadsheets, task boards, emails, and text files. Basically, I'm trying to map out what I think I can fit into the schedule and budget before funding runs out.
It's a bit daunting, almost morbid, to be planning the rest of a project's life. Sort of like when a retirement planner asks you, "well, how long do you expect to live? And at what age will you stop working?" It's a heavy question, but absolutely necessary to plan.
Anyway, that's sort of what I did today. And I think I've got a nearly complete picture of the stuff I need to add, stuff I'd like to add, how long those things take, and about how expensive they'd be.
Financially, I think I'm not too surprised, and not in terrible shape. Far from having everything I need, but there's enough to get by and make a small core game experience. I think I'm a bit surprised at how much of it is just plain old game programming. I kind of expected more graphics programming (which is maybe a week's worth of stuff), and a higher amount of content.
However, I still have to tweak some formulae to account for multiple copies of stuff. E.g. I have stats for the cost of creating a new ship tileset, and the financial cost reflects how many copies of this task I'll have. But the team man-hour breakdown doesn't yet do that. I'll have to delve into that tomorrow.
It's a good overall view of what's ahead, though. And it feels good seeing it all in a single place for once.