Iterative development and what's happening in 2011 | EVE Online

Iterative development and what's happening in 2011

2010-07-16 - CCP Zulu

Hey all, I'm CCP Zulu (formerly CCP Zulupark) and as of about a month ago I've assumed the role of EVE's Senior Producer. CCP t0rfifrans has moved on to become EVE's Creative Director. This means you'll be hearing a lot more from me in the future.

After reading the many discussions following the publishing of the Council of Stellar Management (CSM) 5 meeting minutes, I wanted to write this dev blog to clarify some things that have been floating around, rumors and half-facts that seem to have taken on a life of their own.

First off, Incarna

(9 teams, approximately 70 developers)

Development of Incarna is something that CCP has committed to for multiple reasons, but the most prominent one is fairly simple.

To achieve our goal of EVE being a complete sci-fi experience we must have full body avatars. When we talk to people who like the concept of science fiction games but aren't attracted to EVE, it becomes obvious that there is a need gap that can best be filled by having a character to associate with rather than a spaceship, and I would venture there are many existing players who want this option as well.

As of August 17th, there will be about nine teams working on Incarna. Of those, seven are on loan from other projects - to which they will return once Incarna is launched - and the remainder are EVE development teams. The EVE teams can work on both Incarna and in-space features, but they are assigned to Incarna for now. Actually one of the teams is spending one and a half releases on Incarna, the other about three releases. So the resource cost from EVE is very contained while the size and impact of the feature will be massive.

Dust 514/EVE link

(1 team, approximately 7 developers)

We‘re also committed to producing and launching an MMOFPS hybrid. What we‘re doing is unheard of in the gaming industry, namely linking together a console FPS (MMOFPS) with a classic PC MMORPG. Let that sink in for a minute. CCP is the first company in the world to do something like this. First. Ever. The „resource cost" on the EVE side for that is a team continuing on development of the Planetary Interaction feature (the primary link between the two) for the next three releases, delivering immediate, meaningful value into EVE Online and making game development history while they‘re at it.

In-space features

(3 teams, approximately 22 developers)

Currently, the plan is to have three teams working on EVE (the space parts): a feature team, the mission design team and an art/effects team. The feature team has chosen to work on a PVE related iteration. In short, they‘ll be expanding and iterating on PVE in EVE, bringing together players for immersive and engaging group content. The art/effects team will be finishing up on left-over work all the way back from Apocrypha and on some level even Trinity, where we have not fully taken advantage of new features of Trinity 2.0. This will mean prettier effects that perform better, especially in fleet fights.

EVE Gate

(1 team, approximately 10 developers)

We have one team (at this point) assigned to EVE Gate development and iteration, and that won‘t change for the foreseeable future. Note that this team is a dedicated Web development team. I had planned on using them to fix fleet lag but was talked out of it.

The others!

(4 teams, approximately 15 developers)

I‘ve now accounted for 14 teams, but we release plan for 18 teams. So what are the rest doing? Well, we have a team dedicated to testing automation. Currently they are focusing on "fleet fight in a can," which is a borderline sentient AI-testing automation feature that allows us to instantly simulate interesting aspects of massive fleet fights, getting tighter feedback loops to the expert team that is working on improvements on that front (we are talking MIT Ph.D. here). There is a team dedicated to making tools and reporting mechanism for the customer service department. The core infrastructure team (patch delivery mechanisms, among other duties) is in there, and then there‘s a another team dedicated to performance monitoring and improvements.

On top of that we have some (space) cowboys running around injecting code into Tranquility live to debug fleet battles and fix exploits (these are the same people that do code reviews and architectural integrity checks for the EVE codebase). This is where the MIT Ph.D. come into play. This doesn‘t fall under the classification of a development team but contains computer scientists, QA people, operational staff and others. They are hard-at-work (often well-outside normal working hours) profiling and diagnosing all sorts of lag-related issues or trying to catch fringe cases that are impossible to reproduce on our test servers. This team will be greatly assisted by the "fleet fight in a can" sentient AI.

And that‘s that.

We have an aggressive hiring plan, as evident by our job listings. When those positions are filled we will be able to fully staff two additional EVE feature teams. Finding the right people with the right skills so that they can further enhance something which we have been working on for a decade already isn‘t easy, but we have increased the priority on global recruiting. If you or someone you know is interested in working for CCP, now is the time to submit those CVs.

So what about prioritization? Where‘s the iteration on feature "x"?

We have seen a number of players talking about us moving from new feature development into iteration in the year 2012, but until then we‘ll be pumping new features into the system. This is not an accurate description of how a 200-person development team works. In the 2011 winter expansion we‘ll certainly focus more than ever before on iterating and polishing up all the features that EVE has expanded to in the then eight years. What those specific systems are, I don‘t know as we haven‘t started the planning phase for that yet.

That does not mean that we won‘t make any improvements until then!  We haven‘t started planning the Spring 2011 or Winter 2011 expansions yet (not to that level of granularity) but we do realize the urgency of revisiting certain key game features. Rest assured, the CSM will play a large role in identifying those.

Please, also keep in mind that our developers are constantly fixing bugs - small ones, big ones, important ones and obvious ones. But also bugs you‘ve never experienced, never seen and never heard of. The EVE code base is as big as the universe itself and we‘re constantly engaged in cleanup and backend work that the players will never notice unless something goes horribly wrong.

CSM you say?

Like many of you, I have read the meeting minutes from the CSM 5 meetings in Reykjavik last month. To be completely up-front about it, I have to say I‘m quite taken aback by the difference in the meeting minutes compared to my firsthand experience from attending the meetings.

For my part, I felt the discussions were very constructive and I had great dialogue with the council members.

Maybe I‘m completely delusional and it didn‘t go that way at all, but there‘s definitely something broken in the feedback loop we have with the CSM and I will be working closely with them to enhance our collaboration. When two parties have such vastly different perceptions of the same event, something is wrong and needs to be fixed. We have some suggestions in mind and we will be exploring them in the fall. I‘m confident that with the CSM‘s help we can take this to the next level.


Arnar Hrafn Gylfason

CCP Zulu

Senior Producer of EVE Online