This week I had the opportunity to ask Kihra from the guild Temerity some questions about being a Feral this expansion, and his work on his project – the new combat logging site Warcraft Logs. Enjoy the interview!
Arielle: Hi Kihra, thanks for agreeing to this interview. Can you introduce yourself to everyone?
Kihra: I’m Kihra, a former Protection Paladin who now plays a Feral Druid in Temerity on Windrunner. I am the developer of Warcraft Logs, a new site for combat log analysis.
A: How long have you been been playing WoW? What made you choose to play a Feral Druid?
K: I have been playing WoW since vanilla and have raided every tier, so I’ve been around a long time. I first started playing Feral as my primary alt back in ICC and fell in love with the rotational complexity. I also enjoyed a play style that didn’t involve having to frantically spam a button every GCD.
I was a main tank for Temerity from Ulduar to Lei Shen, but I continued to raid on the Feral as an alt. In Tier 15, I was actually raiding three nights a week in two guilds simultaneously, and so I got a taste of Heroic raiding on the Feral in 10-man. I loved it so much that I decided to make the switch to Feral as my main in Temerity.
A: How do you feel about MoP from a Feral / Melee perspective?
K: I can really only comment on the last two tiers, as they are the only ones I raided seriously as feral in MoP.
I found Feral to be extremely strong for single target damage fights. Feral also ended up being extremely good for both of the “end bosses” of the last two tiers. On Lei Shen, Feral had very high single target damage on Lei Shen, while also having the ability to burst the Ball Lightnings effectively. On Garrosh, Feral does phenomenal damage to the boss and benefits from multiple sub-25% BitW phases because of how the fight is designed.
As far as playing a melee in general, the biggest adjustment I had to make switching from 10-man to 25-man raiding as Feral was the decreased visibility of ground effects because of the increase in the number of players in melee. I am hoping that in Warlords of Draenor they do something to improve melee visibility.
A: Having now killed every heroic encounter in this expansion, what was your favourite as a Feral? Why?
K: I really enjoyed the last three fights this tier: Siegecrafter, Paragons and Garrosh; of those last three my favorite was Paragons. The fight had a lot of personal responsibility and interesting mechanics, and I like fights that challenge you to do high DPS while also being mindful of positioning and of other mechanics that force you to think on your feet.
A: Why did you start building Warcraft Logs? How long did it take before you were able to have people start testing it?
K: The reason was pretty simple. As a software engineer, I was just interested in the problem space. I had a number of ideas that had been percolating for a couple of years about things I would do differently/improve on after using World of Logs for so long.
One of the most glaring examples I remember was the Dark Animus fight, where someone would do something wrong with one of the 25 Anima Golems. Not being able to see specific mob instances in World of Logs made it very hard to quickly tell who did something wrong, etc. So very early on I was convinced that being able to see specific mob instances would be a big improvement. That intuition turned out to be correct (for example, this capability was very useful when progressing on Siegecrafter this tier)
As for how long it took before people could start testing it, it was about 4 months of development before I was willing to give out the client. At the 2 month mark I was logging fights for our guild, and the site was useful enough to test with, but it was another couple of months before I felt comfortable giving the client out to anyone. The site has been in development since May.
A: It feels like there has been an evolution of log sites over the years from the original (whose name escapes me) to WoL to WL. Do you agree? Does WL have a planned lifecycle?
K: You may be thinking of WoWMeterOnline, which interestingly did have some features that WoL did not. I definitely think when you make a new log site, you will obviously learn from the current log sites that are out there, i.e., what they did right and what they are missing. In my case I used WoL (and WoWMeterOnline) for years, so had a fairly good idea of what they did well and what they were lacking.
As far as lifecycle, I hope WL will have a long run! I have a huge list of planned features, so I’m not going anywhere.
A: There are a million and one ways to analyze data from the combat log. Can you give us a sneak peek at what some of the planned features are for WL?
K: The main features I am working on in order to be feature complete for beta are the Expression Editor and Rankings. These are really the only two missing features that keep WL from being a viable replacement for WoL. Once I have those features implemented, I’ll go to beta, bug fix for a while, and then release.
The Expression Editor in Warcraft Logs is going to be very powerful, since you will be able to use expressions to manipulate all of the views and not just the raw events view. You’ll also be able to work with ranges and show your expression results as separate graphs (or have them show up as lines or bands on the main graph).
For Rankings, I plan to have more categories than WoL, i.e., I’m going to have rankings for speed kills and best execution for guilds, and for personal rankings I plan to have boss damage rankings and tank healing rankings (in addition to the usual overall damage and healing rankings).
Post 1.0, the three big planned features are log merging (so fights with phases can be dealt with cleanly), a comparison view (to compare your kill with another guild’s kill of a boss on the same graph, etc.) and combat replay (showing the actual combat play out complete with player/mob positions and raid/unit frames).
A: Are there any features you will not, or do not plan to add?
K: I can’t think of anything I’ve had to say no to yet. Most of the features people have asked for so far have been quite reasonable.
A: What has been the easiest and most difficult part of building a new log analysis tool?
K: The biggest challenge with building the tool has been performance and scalability. I had to implement a lot of clever optimizations both in the packing of the combat log files into a new format for transmission to the server, and then the server itself also has some very complicated caching. I made the decision early on to optimize around viewing a single fight, so Warcraft Logs does that very well (even on considerably weaker hardware than its competitors). Viewing lots of fights (e.g., all wipes on a boss) is considerably slower, mainly because of the weaker hardware the site is running on.
The easiest part about building the tool is that WoL is there to sanity check results. Having results to compare against in order to find bugs has made it very easy to resolve correctness issues.
A: What can the community do to help you make the highest quality tool available?
K: Leave feedback and bug reports on the Warcraft Logs Forums. All feedback – both negative and positive – is welcome. Let us know what features you’d like to see on the site.
A: That about wraps it for this interview. Do you have any shoutouts and/or final words?
K: Looking ahead to Warlords of Draenor, one of the best ways the community can help is to track class/spec detection and stun/interrupt detection. I have posted those functions on the forums so that the community can correct and/or add to them as ability IDs change, etc. Staying involved on our forums is the best way to keep the log site improving and to ensure things you want to see make it into the site!