Chucking 2 weeks' code out...

...never felt so good }:-D.

For ages I was piddling around with a way of displaying the "embryos" - these are enemies that are in the level, but haven't yet arrived on the Web. There can be hundreds of those at the start of a level, and I wanted to display them in a way that looked cool and which would let you "read" them just by seeing them as part of the environment, to see how many were left (and therefore how soon it was till the end of the level) and where they were about to land and such. I had this whole chunk of code which allowed you to rganize them to run on little tracks, and split off from those when they were going to land, and this had got more and more complex and still didn't look that great, they tended to merge and it just looked like a few lines off in the distance and the bit where they emerged and landed on the web wasn't that clear... I was never quite happy with it and it was something I'd been meaning to come back to.

So this morning I thought I'd make a new version (copy my entire source and begin from the copy) and as I'd have a fresh backup if it all went horribly wrong, set about attacking that code and ended up removing most of it. I've now got a new scheme where the mebryos move about in a kind of swarm at the bottom, and there are bright flashes when individuals detach from the swarm to begin a landing, and when they do they leave a nice trail coloured according to the enemy type, and it looks really cool, *and* is abundantly easy to "read" even in the thick of a level where I've got the level sequencing cranked up to 5x normal and a majillion things on the level at once. I was finding I can much more easily rinse the levels with the flowers because I could see where a flower was going to spawn and flit about so as not to kill it straight away but then come back and nurture it, allowing the normal enemies to just pile up and then setting up massive bull runs. It's amazing, it's still possible to play this accurately and strategically even when the level's almost whited-out from so much crap emploding (not that I'll have it that fast or overloaded in most of the levels that the player will see). But now there are so many cues that relate to gameplay events actually embedded in all the stuff that's going on that you *can* read it all and play sanely and prevail. I'm really happy about that. And it feels *so* different from Tempest; it very much has its own feel and identity, which is just what I wanted.

I still need to work more on the tutorial level - the bulling stuff is adequately explained by the tutorial but I could probably add a couple of intermediate sequences that would serve to more gently ease in the stuff about bullet repulsion. And what I'll probably do for the demo version of SG is allow you to play the tutorial level on a variety of different geometries and at different sequence speeds, since geometry and seq speed affect the difficulty. I want the demo player to have plenty to get their teeth into, and to have a good bit of replay value in having opportunities for the player to learn the strategies being taught and then apply them in order to thoroughly rinse each level of points. But still hopefully leave them wanting more enough to drop 400 points on the full version.

Bit like I did with GR++, really, where you could spend quite a long and happy time trying to get as much as you could on the 5-level demo which kept getting harder for as long as you could stay alive. I hate demos that are too short and insubstantial, and I want people to buy SG because they bloody love the demo and want more }:-D.