Hm, ok then. Aparently I wont let myself begin judging(1) the other games if I don’t write a post-mortem first. Heh… ‘post-mortem’ seems kind of apt when you consider the theme.
Without further ado:

What went wrong:

• Dithering with the theme for slightly too long.
  As I said before, my first thought was to make game about Medusa(5), since an ‘instant-statue-stare’ is definitely unconventional, after all.
  Then I found out someone else had the same basic idea a few years back.
  But other than that, the design was quite different!
  Mine would have been a top-down dealy with push-mechanics for mirrors and any now-permanently-petriefied Ancient Romans.
  Still wonder what would have happened if I didn’t check the idea online or just pushed through.
• Not playing with the theme long enough(4).
  I was just so glad I had thought of another workable idea(6) that I started working on it almost immediately.
  Even if I didn’t want to spend any longer on deciding the idea, I think I should have at least spend more time with the design after the idea.
  Also, the first idea was very simple to make, the second required physics (see another point below).
  Furthermore, seeing the end result, the theme may not be immediately obvious to everyone.
• Futzing around with reading levels from a file.
  I did get it to work, and in a time I’d normally consider very acceptable,
  but thinking about it now, it wasn’t nescesary.
  Also, I didn’t have time enough for the extra ‘load your own level text-file’, so the work isn’t even visible to the player!
  More than that, it cost time I could have spent on polish, or level design, wich are the important things this game (still) needs.
• I’m … not really happy with the end result. Oh, pleased to have finished what is technically a complete game, yes, absolutely.
  But to me it seems like it just barely passes ‘good enough to submit’.
  I don’t think there is much you could cut from the game without losing at least another catagory (I already decided to loose ‘audio’ and ‘mood’).
  (Well, the txt-based level load system maybe. But since that was already done on the first day, it doesn’t feel as if I could have cut it.)
• If you’ve got a library, then physics are very doable.
  It’s just a question of wheter you should try to implement it in a puzzler, since testing becomes a lot harder.
  Although my language of choice (LÖVE) has box2d (well known 2d-physics-lib.) built-in (used ‘underwater’ by LÖVE),
  I still wonder how much time I lost fiddling not only with a new language, but also it’s physics sub-system.
  All-in-all, I’m not sure how much time I lost because of this.
• Not being in a team? I can amuse myself on my own just fine most of the time,
  but not having anyone to play these ideas off on made me feel strangely lonely.
  Oddly this wouldn’t have been the case for a larger project I think?
  Though it had definite upsides; I could focus pretty well

What went right:

• There is a complete end result, as in, ‘actual, playable game’.
  Menu, three levels, sprites, sound, win-screen.
  In theory, it might be possible to have fun figuring out the puzzles
  (Well I wouldn’t have submitted it otherwise!)
• I had originlly planned to spend only the first day programming the basics, and the second day on art.
  It didn’t turn out that way (I had only about 7 hours left for the actual art in the end),
  but the constant pressure of ‘must hurry or can’t art’ kept some things nicely focussed.
• Taking the time to sleep and eat.
  I know from experience that taking away sleep-time does not work for me.
  I need my 8 (or in case of the last three nights, 6) hours of sleep!
• ‘Learning on the job’, then still being able to submit about two minutes before the dealine(3), even though I understand it’s … not exactly recommended to:
  • Jump into the compo instead of the jam for a first timer,
  • Let alone doing it with a language (LÖVE/lua) which you have, say, only used lightly for about one or two days before(8),
  • Or carefully picking out audio and visual tools, have no time during the warm-up weekend (or before that) to practice in them
    and then, dring the compo itself, not touching any of those and just using bog standard Windows Paint(2)
• Being a first time participant, I definitely ‘tried something new’ 😉
  (Well, also the language and some other stuff.)

I anyone still wants to have a go at it after reading all this, here it is.

0] Lua has… changed me. Since when did I start numbering things from ‘1’? My programers-licence will surely be revoked if I keep this up! 😉
1] I know there are 21 days, but I also know if I start thinking like that I’ll end up with no judged games one day before the deadline.
2] With a little help from The Gimp to create a translucent background for the sprites.
3] Ok, there was something called packaging hour, but it still felt good
4] That’s the opposite of the first point, but bear with me here!
5] It was going to be named either ‘Gaze of the Gorgon’ or ‘Petrified’. Now that I think about it, ‘Hold Still’ would have been funnier 😉
6] For the record, that is ‘create your own weapon out of parts, then hit ‘simulate’ to see it it actually hits the target(7)’.
7] Gah! I just figured out now I should have made one of those red-white-red circular targets instead of a box that says: ‘goal block’,
dang that would have made things clearer for the player.
8] … but I encountered no mayor hurdles or made bugs that had me held up for more than, say, 30 minutes. Maybe I got lucky?