Might as well post some trials and tribulations here:
What went right1) I finished a game in my spare time. OK, it's only teeny tiny, but still...
2) I was able to implement (if not exactly explore) a game mechanic I've been musing about for a while, and picking this limited scope allowed point 1 to happen

3) I got to play lots of other cool entries and see some other relatively hardcore sci-fi fans come out of the woodwork. I like that.
4) My toolkit/library is growing and I got to use some of the entity component stuff I've been mucking about with for a while. It did make things easier!
What went wrong-1) Not enough time. I didn't start anything really until the flight to GDC, where I took some raycast code I'd layered on top of Box2D and tested the scanning/sensor return mechanics (about 3-4 hours?). Then GDC happened and not much work got done (2 hours). After that I was in Berkeley for a week working for a client, but got work done in the evenings (6-8 hours?). I did a bit more on the flight back (2 hours) and tarted everything up when I was back at home (4-6 hours). For a total of around 24 hours.
-2) Too dry. Indulging my inner sci-fi geek meant not giving instant feedback on correct/incorrect contact identification, and not dressing things up as more of a game. So I'd argue the result is interesting but not exactly fun.
-3) No screenshots for ages, still no video or readme. My thread didn't exactly draw attention, because it was pretty boring, whereas my Space Effort thread was a good source of encouragement for me perhaps because I started with arts not codes.
-4) No sound. Again. Sorry

-6) No UI library. I wrote some IMGUI stuff for Space Effort but decided not to use it again, and my UI this time was minimal so it wasn't a huge problem. But it did mean that things like the difficulty select and play again screen got left until the last possible minute.
-7) It's only tangentially cockpitty. I've been meaning to add a bitmap screen border to make it look more like a display, but didn't get to it in time for the compo version.
Grey areaNo Photoshop on my work laptop. Doh! I was sure I'd installed it too... This would have helped the early boringness issue, but at least this it me concentrate on the game mechanic, so I don't know if this is a net win or loss.
What's next?More interesting (i.e. not just one convex polygon) ship shapes.
I really want to turn Sublight into a proper 2D tactical space combat game. I'd quite like to do a 3D game, but frankly I've been there before and it does make things more complex. I think that most of the same ideas and mechanics can be interesting in 2D though.
I do like the idea of doing the raytracing radar mechanic in 3D - instead of a 1.5D scan section of a 2D ship, you'd get a 2.5D scan of a 3D ship, which you'd draw like a heightfield mesh - e.g. like Toastie's nifty holomap in Landfall.
Hope the above was of interest, I like reading other peoples' post mortems, so hopefully someone else will get something from this!
Cheers and thanks all,
Will