Tuesday 10 May 2016

Successfully finished an alpha in time for the dev contest. Will need a bit more clean-up still, though, I have to admit.

I took a week away from the computer after the insane crunch to help stave off burnout and I think it was successful but just to make sure, I'm gonna spend some time working on a little side-project I cooked up with Bethan. I call it "SNEK" and it's basically a rather strange Snake clone with a few odd twists - most notably, the game stored the entire set of moves, allowing me to do strange things with time.

Current plan with that one is to cook it up in HTML5 and pop it on a website. Might even embed it here just for the oft-mentioned shits and giggles :)

Friday 29 April 2016

Furious Coding Update

I decided a little while ago to enter Gestalt into the WG Labs Dev Contest, which ends tomorrow evening. I kinda realised midway through last week that the closing date was rapidly approaching so I sat down and worked out a pathway to Alpha, and it came out as 56 bullet points.

As of this evening, there are 20 left.

Agonisingly close and yet I'm not giving up just yet. If I can get it into a state which can at least be called an Alpha then I can still enter the contest. So for now, insane crunch GOOOOO!

Wish me luck and I'll see you the other side.

Wednesday 6 April 2016

Quick update because I don't post here nearly as often as I perhaps should.

Progress is still being maintained. I've decided to try to reach Alpha by the end of April in order to enter a contest, and have cleared out (nearly) all of my bug list in order to enable that. I have the outline of a plan to Alpha ready, so I just need to knuckle down and get it done.

Part of that plan is to start implementing the structures necessary to permit multiplayer, since that's something this contest requires. As can be seen from my old posts about Gestalt, multiplayer's always been on the cards, but it's not been forefront in my mind so I think I have some refactoring to do to make it work. Probably a good idea to do that now, tbh, even before I have a single line of netcode.

Finally, gonna try to get Fraps to give me some decent video from the game. Will post it up if it works.

Monday 7 March 2016

Node connector points calculate nicely and without unnecessary overhead (my first pass on it was embarrassingly poor, so I immediately redid it). Worked first time, too - I always feel a bit suspicious when that happens :D

I also turned on Box2D's Continuous Physics mode because the change in physics scaling has led to some rather unfortunate tunnelling issues with the axons and the bucket. It has had a really crippling effect on the framerate, quite pathetically so in fact - under 30, in a game that consists of nothing but circles and lines!

So I set up Release Mode and was gratified to get a stable framerate well over 120. Well, that's something at least. Picking still doesn't work right, I need to completely remove the ScrollComponent and bake my own replacement (which should hopefully fix pinch-zooming in Android and the weird scroll wheel problem under Windows, as I think it's ScrollComponent that's causing both issues), and that's a big job, so with the school run rapidly approaching I've decided to leave that for next time.

Still, though, a hell of an afternoon's work.

Thursday 3 March 2016

Refactor coming on nicely. I now have the map and the node contents rendering from the same component, and the modelling classes all working strictly in physics space (which I've decided is object space for this project). I've put in a simple "select to view" control method and it's now possible to "look inside" the nodes to see what's there.

It's a bit scrappy looking at this point, so no screenies yet, but very soon.

Tuesday 23 February 2016

I've lately been working on Gestalt again (on the basis that it's the project most likely to produce a working product in the next six months) and I ran into a very sudden realisation.

I was facing spiralling levels of difficulty in implementing new functionality. In particular, translating the mathematics of the mechanics from the physics model in Box2D to the screen component used to render the inside of the Node was becoming especially complicated, particularly given that since I discovered that running multiple simultaneous Box2D worlds was impractical, all Nodes had to exist in the same physics world. While I could probably save the previous state and restore, it would be impractically complicated and lead to weird artefacts.

Then somewhere along the line while trying to implement the "flying along a node line" functionality, I realised my error. Having two separate view structures - the Node Map (which now corresponds to the physics world) and the Node Screen was the core of the problem.

Then I realised what was happening.

The existing "Gestalt" is a tech demo, building apparently-living "Axons" from simple physics objects.

But I'm trying to turn it into a game. And that involves a fundamentally different approach.

Quite a large refactor coming up. May be a while, especially given my rather cripplingly slow development rate and a couple of other projects on the go. But once it's done, it's very promising indeed :)