In between the hassles…

Taking breaks from trying to solve the game’s problems and actually playing the game (as my non-caps name) has keep me running these last few days. I’ve been playing BioShock too, but far less than our game.

Watching the lightbulbs going on over players heads as they get to grips with whats new in 1.27 is pretty fantastic. The way people conduct their battles didn’t suddenly and magically change over night – but it shifted into gear.

The first pass was people driving supplies around willy nilly, which doesn’t work because trucks are kinda soft targets. However – going up against an enemy with supplies is a different beastie than going up against an enemy without – a battle between two equal forces is going to be decided by which co-ordinates with its suppliers better.


Well, individually, as I’ve said before, they’re zilch. Like as not, you’ll go thru your standard ammo issue just ranging a target. It’s also surprising how short 500m is – especially for a round that takes ~20s to travel the distance.

But the mortar is anything but useless. I didn’t imagine myself spawning it more than a handful of times. Oh, silly me ;)

No Austin for me

I was all booked to go to this year’s Austin Game Con. But up until yesterday we were fighting with an old bug that inexplicably chose Sunday to surface. Really, this bug has been in the code since Thunder was working on the host with me – over 2 years ago. The baffling thing is that the code is in a critical path. It has been executing every minute of every day, for every player connected to the game, for nearly 3 years. And then, on Sunday, it stopped sending an 8-bit value of “0” (zero) and started sending values between 1 and 16.

How the hell?

At the same time, it highlighted a shortcoming of our QA systems that allowed us to put different terrain versions up for the game servers and game client. Sometime, long ago, there was a good reason for this, but nothing could have been good enough to justify making it the default behavior of the QA pipeline. When Goph gets back, we’ll get it integrated into the QA pipeline.

The first bug presents me with more of a problem. Our network API is actually rather good, its designed to avoid issues like this but this was a clear cut case of bypassing those measures; I can modify the bypass mechanism to include checkpoint markers.

Depending how things go tomorrow, I might still fly out to Austin for Thursday and Friday.

I could use the break.