Well I have (more or less) successfully componentized the npc ship AI. This is great news, and makes it at least possible to have more interesting attack and behavioural patterns, although right now it's very very limited (just one AI module, and only one implemented for it).
One interesting note: I used basically the same system for generating the models (e.g., flying enemies, whatnot) that I did for this. If that strikes you as amazing, well it should! When I say system, I really literally mean I used the exact same class without modification to get the core part of the work done. I need to make some kind of theory on the kind of data abstraction I'm doing in vex becuase it is very useful and would be nice to formalize/flex it out a bit.
So I'm not sure if, for the sake of gameplay, I ought to continue improving this (npc ship AI), or if I should work on something else. It is nice to take a break from this kind of coding because it tends to be sort of tedious, and lots of annoying debugging.
There are a few obvious candidates looking for attention: I could code weather stuff, I could code stuff for ship engine-fire or other effects, or I could put in code for autogenerating weapons.
I think I will put the code for autogenerating weapons. Frankly, it's not too much work and it's one of the, if not THE only major thing that still isn't autogenerated.
I should mention that I had a great idea to really really help me meet this impossible production schedule. Basically, I have to "flesh out" a lot of the autogen models by creating more seed models. Just like I need to flesh out the dialogues by adding more and more dialogue branches/groups/etc. Each work product is small -- taking literally only a few minutes total (maybe 5 or 10) but I need to create a LOT of them to get a good effect. Before, I had thought, what I will do is focus on these once everything else is done. But now I've decided it makes more sense to work them in lockstep with the rest of development. That is, when I'm waiting for a compile, I go and make another seed model, another texture, etc. This could ultimately knock a huge amount of time off the schedule, which only makes it more possible =) Anyhow, I am thinking I should work with some kind of quota for how many of these elements to create each day, in addition to the other primary goal I am interested in meeting.