sen4mi

This sounds like a missing dependency - or maybe vortex refusing to replace a file? But to solve that, I imagine you would have to go over each FG body related file to make sure that the version is deployed is the version that was supposed to be installed.

In game, no approach happens, even when manually trying to trigger an approach and apparently suitable NPCs are nearby.

Approaches sometimes have had a problem for me, where when it tries to find an NPC to approach it grabs an NPC which does not exist. I do not fully understand why.

Hypothetically speaking, you could set iXPRewardKillNPC___ and iXPRewardKillNPC___ to zero. If I understand correctly, this would prevent all experience from killing anything. You would then have to arrange things so that all such NPCs would be off the board when you switched back. (And that all other NPCs would remain off the board until then.) It might be possible to intercept and replace the routine which awards experience, or to set an on-death effect which awarded you negative experience when you killed them, but getting those right sound like extensive work.

In addition to SF Extended Skeleton, Haters Male Body Replacer is basically a repackaged (with permission and encouragement) Roberts. That said... I finally managed to see how this mod plays, and ... well, ... I can see how it would be great for some scenes, but it is very much not what I was hoping for.

Games exist out there, where progress stalled for a long time, but the game did not die from it. (Doom 4, Team Fortress 2, Final Fantasy 15, Metroid Dread, Alan Wake) And, there's mods which had similar issues. (Black Mesa, Enderal, Long War 2) So, um...

I would like to make sure I understand this concept of "scene events". Are these meant to be events which are triggered during the progress of the scene? Or are these meant to be events driven by the user and/or papyrus which can modify the scene? (Or both?) I am particularly interested in making events from the user a part of the scene. These could be "struggle events" or they could be something elaborate (perhaps choosing between various alternatives and/or something like switching to a different scene). But I am worried that that's not what the RFC is talking about. (Of course, these things are already somewhat complicated because of the limitations of the animation mechanisms - though perhaps physics and/or keyframe interpolation could make that pain bearable?).

Thanks, I guess I'll see how this works when I get to that part of the game.

Unfortunately, in my instance of starfield, the new Hilda Fournier is headless. (Possibly because I am using zbb? Usually zbb just causes clipping, but ... I don't know.) Also the conversation in the restaurant, before meeting her, explicitly stated that she was male.

Or, generally speaking, artificial intelligence represents "the beaten path". You should not rely on AI to perform logical operations. You should rely on AI to report on well understood issues which happen to be new to you. You can get seemingly conflicting answers from an AI by hitting it with the same issue at different levels of abstraction. The point being that without real world experience to guide it, constructing an internally consistent model of the real world is incredibly difficult. The same word gets used to mean different things in different contexts, and only through heavy repetition would an AI pick up on when the contexts match sufficiently for the words to mean the same thing (or, perhaps, on when seemingly the same thing was actually supposed to be different things - it's up to you to make the judgement call on whether or not the AI got it right).

This kind of bug is probably caused by either (a) not counting the relevant equipped devices (or not scanning the equip slots for relevant devices - either approach works), or (b) not having a fallback count/test event that happens on game load and/or cell change. That said, note that there's scripting issues with how fast stuff needs to happen, so it's also a good idea to chain work across multiple frame refreshes if possible (which means implementing a little state machine whenever timing isn't super critical, rather than using a regular loop or something worse, like recursion).

Well... yes and no. I had not specifically tried that approach, but I had in fact tried that approach. Sorry, the question is a bit funny - I have been using a batch file named "CLEARYOURDAMNCACHE.bat" which deletes that cache (among others) and I have been running it occasionally, but I had not run it in the last week or so, so I had not tried it specifically in the context of this problem. That said, I tried it just now - but it did not fix this problem. Still, thanks for the suggestion - this seems like good information. (Edit: it seems very likely that many of starfield's reliability woes stem from inadequate cache management - it seems likely that it's relying on cache entries which should have been marked invalid by some part of the code base.)

Isn't outfit studio about adjusting a mesh so that it will fit properly on a body, rather than coming up with morphs to be applied to a mesh (the body and the outfits, both)?

Maybe https://www.nexusmods.com/starfield/mods/17434 would work for you?

Remember that the game only supports 253 large esms.