EVE General Discussion

I don't know what's going on inside of CCP, but short of arbitrary deadlines and expectations of progress that aren't being met and bad quality control, here's a few things you should think about and do, if you don't already:

A) All releases should be flexible. If the release isn't good, don't do it, postpone it until it's ready. Player trust and respect is not an infinite resource and every broken update taxes their patience.

B) All releases should be reversible. If the release is bad, roll it back, fix it, only release it when you're sure. The cost of a bad release should be an incentive for people to pipe up and say 'no, if we release now, it's going to cost us'. If anyone thinks a release isn't ready, it shouldn't happen. None of this late night burnout hacking to fix it in time for the next downtime- you'll burn out your employees that way.

C) Test, test, test. Mirror Tranquillity regularly and deploy the entire final release there (in the software industry, we call this a release candidate). Have a mass test to get people playing around with it. Most of the bugs your players ***** about on day 0 of your new release would be caught on Singularity fairly quickly. Once caught, fix those bugs, roll a new release candidate, mirror Tranquility, mass test again. Keep doing this until it's ready.

D) In-house testers. One of the greatest false economies in the software industry is not to have in-house testers. They're relatively cheap compared to your experienced employees and free up whoever is doing your testing to actually spend their time doing whatever it is they're paid to be doing- testing isn't worth their time until a problem has definitive reproduction steps. QA leads can also keep their eyes peeled for the 1/100 talent that ends up in your tester pool.

E) Be aware of your reputation. Unpopular features, for better or for worse, that's the domain of your designers and you have to ride that one out as long as you trust in their process. However, bad software process is entirely avoidable and CCP has a bad history on that front (yes, I know, it's negative bias- we notice everything that goes wrong and not all the things going right, but that's what makes it so important to be rigorous and thorough and minimise the number and severity of what bugs do make it into releases).

F)Be honest, communicative and apologetic. We're all human, we make mistakes. Just be aware if you keep making the same mistakes, you will lose good will (especially from people who know that you should know better).

Again, I don't know how much of this stuff you do, or how well you do it, but obviously there's room for improvement. Maybe a Devblog giving an insight to your development process is long overdue?

Its amazing how many software developers play EVE

not really in fact, eve has always gethered a crowd of computer tech ppl, and many of those ARE actually, devellopers, or do work with some.


unlike many other game, i'd say the average eve player has a better understanding of computers, and the % of ppl working in this field is way higher than any other game

actually, if it would be the case, i'd rather see them keep the 12 weeks pace and actually improve the quality of each release, by giving it more testing, and a better integration of player feedback.

this would avoid coming up with things like the ESS or the odyssey exploration, and bugs like the standing thingy we have atm.....

Since you took the time to write this, I will give you as frank a response as possible on the subjects that are under my control.

C - We were on Singularity from the 6th January with this release. Familiar as I am with the term release candidate, being in the software industry myself, I can tell you it was deployed on Monday morning to Sisi, as is par for the course with our deployments. This might seem crazily short to you, but our development method allows us to work right up until then because we do not hamper teams with code/feature freezes and have mature teams which practice 'Whole Team Quality' (more on this later) giving them the responsibility and accountability over the quality of the features they ship. This is a process I have seen evolve in my 2.5 years here to really empower teams, and it is one that I think works well. Further than this though, we deploy daily updates to Singularity to ensure users have the latest stable build available. Combine this with the internal QA coming from the teams, and our outsourcers running a full regression suite every 2 weeks, and you're looking at a pretty robust testing process.

D - I assume on this you are referring to a test house of entry level "testers" working for minimum wage with a 50-90% bimonthly churn. This is firstly unfeasible in Iceland, as the pool of potential employees does not cover the churn that these outfits have. Instead of throwing a large amount of un/semi skilled workers at the problem, we use highly trained and experienced outsourcers to handle our regression testing, and assist with processing bug reports, providing focused testing at team request, and performing targeted exploratory testing (the most valuable type of testing, for me). Our current outsourcing partner has been with us for 3 years now, and I travel to them each year to ensure their methods and practices are keeping to the high standard established. We also have a proven conduit of taking talent from them directly into our own workforce. CCPers who have taken this route include Legion, Sledgehammer, Lebowski, Antiquarian, and even my good self!

Devblog - As the current facilitator of CCP Reykjavik's Agile Community of Practice, I tabled that we write exactly such a devblog not a month ago! We do need to get started (I actually wrote part of it around a year ago, but it got tabled at the time) but there is a plan in place and many keen contributors.

This really doesn't help understanding how such a bug can see the light on TQ, CCP Goliath. Your post just makes if look worse...

Its not a bug, its a feature and a great troll :)

That's all really nice but doesn't change the fact that VERY often the most hilarious bugs or issues simply make it through testing, SISI and bug reporting, obviously something isn't working as it should. Minor issues can always happen but CCP really has a vast history of royal fuckups, facepalms and "seriously?" in this regard, ones that if QA would actually play the game a bit (and have a clue on what they're doing) simply wouldn't happen.

What's the chances you have "CCP CAPLOCK" type of clowns working in QA?

No boot.ini issues? Ship it.

The inability of CCP to scrap crap game designs that don't sit well with the player base has always been one of their major downfalls.

Anything less than a player riot in-game and CCP keeps their heads buried in the proverbial sand.