Buckly posted in Discord, so I wanted to let everyone know in case somebody missed it: Buckly announced a hiatus yesterday. May he rest in peace, and then come back rejuvenated with more energy
I have volunteered to take over Progress Updates (although I’ll pass that duty on if someone else wants it), but in the meantime we have a bigger decision. Buckly was our entire game design department, in charge of making balance patches, and approving or rejecting changes that effect gameplay. Now we’ll need a new process for making those decisions, and in my opinion the microbe stage is getting mature enough to start asking some deeper questions about how the parts fit together anyway.
The two easy solutions I can think of would be to appoint a new gameplay lead, or have some more formal way of voting on things. Personally, I would NOT suggest we just talk about things until we get consensus, due to the fact that that hasn’t worked so far and tends to just favor never changing anything.
I’m new here, so I don’t know anyone well enough to nominate people, but every team should have a vice lead for times like this, especially in an open-source project where taking a leave or sometimes just vanishing is very common.
Just a heads up that while I may be going on hiatus I’ll still be around if anyone needs me. I’m not quite gone, just distancing myself a little so I can get some breathing room and focus on some other hobbies for a time. I can’t help it I just can’t take my eyes away from Thrive, help. I do agree that the game design team needs to have at least one other person though, which would help tremendously.
But yeah, don’t be afraid to @ me if you have any questions or need something to be "ok"ed.
One huge problem is Github PRs that have an gameplay (or even game balance) impact. For those I’ve tried to ask on the dev discord or buckly directly to check them. Would opening a new thread where I can post whenever a PR needs gameplay or balance feedback help?
Any forum message will auto post to discord and also if someone doesn’t visit here often they’ll get a summary email, though that’s probably always too late to actually test anything.
We need some kind of solution as I’d really like to improve how long it takes to accept a gameplay affecting PR because in the end often I’ll just relent and go “it probably doesn’t make the game too much worse” and merge once the code and bugs are solved.
I imagine we will have plenty of folks abstaining (or just absent) for each vote, and we would want to keep each poll up for a few days anyway to make sure everyone gets to see it and think about it.
My guess would be that we aren’t expecting everyone voting on a gameplay direction to have sat down and tested it thoroughly. To get testing done, I have two suggestions:
Make a written backlog of everything that needs a tester somewhere. I think for now we need to give up on trying to get everything tested in 24 hours (as you know it usually takes a week or more), so maybe we could get more done if testers come online and say, “what can I test right now?” instead of a ping going out and everyone independently thinking, “I don’t have time right now, I’m sure someone else will handle it.”
Have folks who aren’t doing much else volunteer for a week of testing service. I think it’s becoming clear that we have a shortage of dedicated testing staff most of the time, and it’s a well known problem that, with a large enough group, everyone starts to assume someone handles it. Maybe if every Saturday (that’s when PUs go out) we just ask, “who would have time to test things this week?” some people will think of themselves as in front of the line.
Those sound like doable ideas. How would we concretely implement them? I started immediately thinking of implementing a sort of queue system in ThriveDevCenter with the Revolutionary-bot commenting the status on PRs and also being able to be mentioned to request testing, but that would take quite a while to implement entirely. Would we just maybe make a separate kanban for testing team to see which PRs are waiting testing? Like I created for Graphics?
If we want to go with a more lose way to test things, to give community members access through the launcher to be testers, that would also be quite a bit of work. I’ve had a task open about making a public beta feature to the devbuild system. One huge drawback to this is that most new testers are almost useless for the first few PRs they test as they don’t know how to focus on testing it properly and may report bugs / changes that have been already merged to master while they weren’t looking and waste review time going through them.
As much as I like getting code in front of the player as quickly as possible, I think trying to deputize specific community members as testers and get things set up for them will just bog us down more. Personally, I like the Trello idea if Graphics is already doing it, but I had wanted to wait and let someone from the testing team chime in.
As for polls with gameplay decisions, I’m going to take on running an experiment after the release. The topic will be on flagella, since there seem to be a lot of murmurs in the team about buffing them, so hopefully we’ll get some engagement.
I agree. I won’t be able to cope with the amount of tech support I’d need to perform to make that workable…
@RacerBG is like the only member, so…
Also the graphics has just a different kanban board on Github, so it isn’t that difficult to setup. Graphics
I think this is a good idea.
Quite often a PR gets bogged down because we can’t do proper A/B test on it to see if it is actually better or makes things worse (this is always my worry, which is why it might seem I’m blocking a lot of PRs from getting through quickly).