This morning I thought I'd clean up Fleets a bit. Is now the time for that?

As usual, what happens today isn't what I planned. I recognize a design issue, implement a very interesting test, then generalize what happened to consideration of real teams. Watch how my thoughts evolve.

We need a GameOver object and a Quarter object. Let's try one or both of those.

I think there's just one more smart Fleet to remove and then we can greatly simplify Fleets and its interface. I wonder what other specific Asteroids notions may be lurking in the upper levels.

I think I'll move to replace another Fleet subclass. People I trust are tired of Asteroids. I briefly address that concern.

While I believe that my preparation work the last few days has been valuable, it doesn't feel much to me like real progress. Today I want a better feeling. And I get it!

Some thoughts on tests and code, inspired by shipping a defect and by ideas from Rickard. And then some changes. I have a nice one in mind.

I think I can do better than I've done with the Fleets object. Let's think about it.

Let's start with a question from Rickard, and then have a look at the Fleets / Fleet setup, to find one or more small steps toward our desired design. ALSO: Big mistake! Released broken game!

Let's see if we can reduce accesses to the specialized lists in Fleets. As so often happens, the next step isn't in the direction I expect. That's OK.

I think I'll start by counting missiles in the Ship and probably Saucer. Then we'll see what's next. The thing is to keep stepping toward the larger goal. Some interesting [mis-]steps.

What additional special objects do we need, and how might they work?

Let's put the full interaction protocol in place, so that more of our special Flyers can operate.

While sitting in a mostly boring on-line session, I'll do some small safe changes.

I am going to do a thing, and then another thing. Don't tell Hill about the other thing.

I am oddly without a plan this morning, even my tentative sort of plan. Be prepared to read my sort of thinking. I hit a few bumps but not major ones.

We make more small changes in the direction of independent objects interacting to make the game emerge. Let's experiment with full interaction.

I'll do a small change, then see what comes to mind. As long as we keep moving in roughly the right direction, we're good. Thoughts on small steps.

As I move toward the decentralized model, I need some thinking about next steps. But first, an interesting test.

The double dispatch is doing its job. Let's see about cleaning up the references to fleets and whatever else comes to mind.