Ramblings from Tux | EVE Online

Ramblings from Tux

2007-01-09 - 作者 CCP Tuxford

What has Tux been up to?

Well, what do devs do after a big patch like Revelation? We look at the sorts of issues that crop up after the patch, we look over the features we put in and we make sure they're working all right. A lot of stuff needed to be done for this one, but in the grand scheme of things most of them were rather minor - at least the ones that I could do anything about.

Naturally we've been playing EVE a lot, since we want those new toys and features every bit as much as you do. I've been spending most of my time scheming to make some ISK so I can upgrade my ship. I've also been yelling at people and telling them to build me rigs although that is all looking a lot better now.

One of the problems we had was the low drop rate of rig components. The thing is, we don't really have any direct control over the price of rigs. If we want to make a new ship, we can basically decide what it's going to cost. If we want to make a battleship that costs about 150 million, we can just make the material cost roughly 1.5x that of the current tier 2 battleships. This was a bit hard to do with rigs as the material that is used to build them didn't exist until we released the rigs. That means that the only way for us to "balance" how expensive they are is to adjust the drop rate of those components. There are a lot of variables that enter that equation: how often stuff spawns, the chance that it drops this item, and the stack size of the drop, to name a few (or maybe that's all of them). As you might imagine, it is a bit hard to estimate the price of the rigs from drop rates. Given time, though, we should see the price of the components and rigs stabilize.

Booster production is another thing. The manufacturing process is complex; since most of the components needed to harvest the gas and manufacture the booster aren't easy to come by, there are lots of places where bottlenecks can form in the pipeline. Right now the biggest bottlenecks seem to be the skill and for harvesters. Skills are never deleted from your character, so they don't have to have a very high drop frequency. The harvesters can be destroyed, but you don't really need that many compared to the booster production it supplies, so the drop rate is a bit low. Both these items can be increased in drop frequency.

I've been looking at some code recently as well. I don't really code, but that doesn't mean I'm not qualified to do so. I looked at the issue with cynosural fields and cloaking devices. This bug is a perfect example of how a seemingly trivial shortcut can come back to bite you in the ass later. You might think at first that we don't really care about this bug, but we have actually tried a number of fixes. Some have simply not worked, while others have broken stuff more badly than this bug did. I can say with reasonable confidence that it's now been fixed internally, and we managed to not break anything.

Aside from that, we've been talking about what we want to do for Kali 2 and trying to prioritize our ever-growing task list. As I think I have mentioned before, and as Nathan mentioned in his speech at the Fanfest, the focus for Kali 2 is going to be warfare. This involves getting capital ships into the fight, splitting up the blob, improving POS warfare and more. That doesn't mean that other issues and tweaks that need to be made will be overlooked. I'm not gonna tell you that everything is going to be "fixed." First off, not everybody agrees on what is broken, and secondly there's never enough time to do everything that we want to do.

OK, I lied to you about warfare being the focus. The actual focus is on speed and optimization. The most obvious improvements we can make are optimizing the code and adding more and better hardware. So how does that involve content developers? You've already seen a change that helped improve lag, back when we reduced the number of drones from 10 to 5. Similar changes can be made with NPCs, such as reducing the number of NPCs and increasing their strengths, bounties, loot and salvage materials accordingly. To be honest I think we should do that regardless, as I often find it a tad silly that with a single ship I can warp to a complex and clean out something like 100 ships in total - in an assault ship.

Other stuff we can do to increase server performance is to lessen the load on hub systems like Jita, and reduce the blob warfare, which really ties in with the whole warfare focus. Now don't panic just yet; we are not going to sacrifice gameplay for this. All this stuff is something we want to do, and it just so happens that it will help improve server and/or client performance.

Well, thats it for now.
Best regards