Today started with some good ol' bug fixing. My first order of business was fixing some NEO Scavenger bugs. It looks like the locked/disabled iSlab allowed players to tell time, and Rollo's conversation encounter lacked an exit choice. Both of these were pretty easy fixes, so I knocked those out in a few minutes.
Next up was the problem I described yesterday with the neko and Windows targets of the social prototype. It took some considerable trace-debugging to figure out, but I eventually narrowed it down to three causes.
Hey Folks! Hope everyone had a good weekend. It was melting hot here the past few days. 40C on Saturday, and 42C yesterday, with no wind. Being outside was unbearable, and indoors would've been worse were it not for a valiant window AC we have in our living room. Poor thing could barely keep up, and even with it running, the house was up around 27-28C.
We had some merciful cloud cover today, and the temp was "only" in the 30s. It certainly got me thinking about what the world would be like if this becomes the new norm!
I spent some more time testing and tweaking the social stat sim today, and noticed one potential issue. I set the AIs loose while I was off eating lunch, so I could have a long data set to plot. And the result revealed some bad trends.
Overall, it looks like the stats have a prevailing trend away from zero. Some decreased linearly over time, others increased. And while some of this can be accounted for by the gradual "tick" functions which increase/decrease each stat over time, some of these stats were affected by greater forces.
Continuing yesterday's work on the social interactions, I decided to do some data analysis to see how things were trending over time. It seemed like the discreet interactions were doing what they should, but what's the big picture?
Well, as it turned out, there was a pretty major (yet sneaky) bug in the code. The game was incorrectly loading interaction data for the effects applied to the granter (recipient) in an interaction, and it was causing weird drops in stats. This was fortunately easy to fix, and once done, this is what life looked like for one AI:
I managed to get the latest crew interaction prototype changes working.
To recap, I recently decided to change each interaction to have effects on "us," "them," and to also have a list of expected effects for "us," to help each crew make decisions. The crew looks at each available other crew, checks each possible interaction for a "score," and chooses the best fit for its current needs.
Hey Folks! Today's work was a mix of business admin and prototyping.
Hey Folks! Hope everyone had a good weekend. We had the most perfect weekend weather here. Sunny, 25C, a nice breeze, cool nights...and we took the opportunity to head into town Saturday for an art show, some food, and wandering. We saw this really neat door knocker made by a metalsmith student. It was a skinless human leg that swung at the knee and hit a tree trunk with its heel to knock the door. Bizarre, but cool!
NEO Scavenger is now officially updated to v1.11! Since the test builds have been relatively stable, I've just finished updating the default builds to 1.11 on all sites. The "test" links are no longer necessary, and have been removed for now.
I was able to fix the bug in the petitioner/granter interaction code this morning. I had a redundant check for a reply going on that was causing the process to exit before it started. It also had a bug that miscalculated goal values, resulting in every goal being a chat reply instead of just a chat.
I also made a tweak to the way the map lists crew when queried, so the list is always provided in a random order. This helped reduce the imbalance of requests on whomever was listed first in the map's manifest.
Hey Folks! I continued work on the social sim prototype today, and made an interesting enhancement to it: petitioners and granters.
Initially, I got the system working such that crew slowly increased social stats, each stat at a different rate. And when those stats crossed certain thresholds, they got discomfort conditions applied to them.
Hey Folks! For those of you who watched, today's interview seemed to go pretty well! We chatted about NEO Scavenger's inspirations, my design process, plans for the series/world, as well as the prototypes I've been working on. And those lucky viewers also saw some exclusive artwork, mockups, and screenshots of the upcoming prototypes! Thanks to all who watched!
On the development front, it was actually a pretty productive day.
Hey Folks! Hope everyone had a good weekend. It was pretty low-key here, and I had a bit of spare time to enjoy the spoils of the Steam sale. Two favorites so far include Creeper World Eternal and The Wolf Among Us.
The former is a sort of tower defense game where the enemy is a contiguous goo rather than discreet enemies. One fights the flow of this goo over terrain heights and channels, and using tools that alter the flow as well as destroy the goo.
Today was a brainstorming, research, and design day. Now that I've got a rudimentary AI loop running, I want to see if I can have it simulate something more complex than a single need replenishment.
My first instinct was to start looking at personality classification. Were there common traits that I could sort out from a larger list? Was there some way I could describe a personality in code based on a limited set of attributes?
We had an internet outage this morning, so I used the quiet time to do some more work on the conditions V2 prototype. I've got AIs running around a map right now, and as time passes, they grow more lonely. Then, they seek out other AIs to cure the loneliness, pathfind to them, and exchange social conditions.
Hey Folks! Hope everyone had a good weekend. It has become super hot outside, and we've boarded up windows to block the sun and heat. My 1920s-era Bladerunner detective fan is pointed right at my face, and doing a pretty good job of keeping me cool!
Most of today was spent getting up-to-speed on the NEO Scavenger tablet project. Steve has been working away on project setup, repositories, etc., and I've been getting my dev machine synced with his so we can collaborate.
Continuing the AI/conditions v2 prototype, I decided to try modeling some social behavior. As a simple test, I've switched all the hunger-related test conditions I was prototyping to be social stats. Things like StatSocial, Lonely, and related thresholds.
Today was the last of my weekly classes, so a good chunk of the day was spent doing that.
However, I did manage to get a bit more work done on the relation prototype. I finished up the data parsing of XMLs (and created some sample XML data), and the prototype is now running using XML data. I think the next step will be to start adding sample emotions and relation conditions, and see if I can come up with a workable model based on those.