Arc, offers, and hiccups

It's weird how I forget about this blog for long periods on end. Quick update on what I've done in the 3 weeks since the end of January:

After getting collisions and keybindings to display, I took a break for a week and ported Arc to JavaScript. It came out in early February, and I was initially going to stay away, but it was just so tempting. I needed a break from UI widgets anyway, and an interpreter is just the sort of quick project that can provide some variety.

'Course, that got some attention, which is good for me but bad for the Diffle. In that week, I got 3 job offers - one from Drew at DropBox, one from my friend Doug out in SV, and one from Paul Buchheit at FriendFeed. Drew also put us in touch with the FuzzWich guys, and we had a long e-mail exchange back and forth. Apparently they were working on a game-creation system before they did FuzzWich, and they have some really impressive credentials in game development. No wonder YC didn't go for us. The sphere is apparently a tough nut to crack: they kept running up against problems about making the games playable enough to be impressive yet easy enough to create. Also, apparently my architecture is insane.

That was probably the closest I've come to quitting so far, between the existence of very attractive alternative offers, and the confirmed difficulty (by people with way more experience in this than us) of the problem sphere, and the prospect of rewriting our whole software in Flex. I mentioned the job offers to Mike, and he said to do whatever I felt was best, so I'm not really tied down to Diffle.

But I'm going to stick it out for now. I'm more convinced than I've been in a while about how this is really a prime space to be in, and that the market's heating up. The big question is whether we can execute on it, and I'm not sure we can, but the challenge of trying is kinda fun.

Anyway, I was kinda burnt out for about a week or so after releasing ArcLite (this was when the e-mails were flying back and forth). When I got back to work, I decided I'd move back to the compiler and update it to meet the full requirements, instead of just the skeleton we've been using to prototype the editor UI. I also wanted to correct some mistakes in data representations that have really been hurting us on the JavaScript side - we can simplify the editor widgets a lot by making the game data structure more sensible. This work has been moving along quite well; I initially estimated 2-3 months until full launch when I first set out, but it now looks like we may manage in half that time, barring extraordinary hiccups.

I'll do a separate post with the current technical hiccup that's been stalling me, for the last few hours at least.