Arena Commander Update: A New Hope
NOTE: (This article started as a news update but became something more. I slapped the Blog tag on it to indicate that part of this post is personal opinion and analysis, and am adding this note to make sure that is clear. - Ed.)
(Citizen Star News/Citizen Ed) - 2014-05-31 - Recent progress reports and chat comments are leading us to believe the critical and blocking bugs holding up the v0.8 release of the (pre) Alpha of Arena Commander (previously known as the Star Citizen Dog Fighting Module) are all resolved. The weekend will spent testing and seeing if any new bugs have been introduced or old bugs reappeared.
UPDATE: Ben Lesnick was a guest of Guard Frequency today. He's passed along from Travis Day that no new blockers have cropped up. A few new bugs, but nothing too bad so far. Go/No-Go meeting is slated for Monday morning.
This will be a critical test of the production process. We keep hearing about "change one thing and several other things break", which is quite concerning. Hopefully in this tightly change controlled environment, this can be kept to a minimum. We could potentially see a release early next week. This would be a relief to people involved with The Next Great Starship as the pressure in the live broadcast with an audience on June 7th would likely be not desirable.
Why is this so hard? Why are people upset about a few day delay? What does this mean about the future?
Arena Commander v0.8, is this an Alpha or pre-Alpha? Why do people say there is no such thing as a pre-alpha? I think it's fair to call this release a pre-alpha for Star Citizen (since it's only a fraction of the intended functionality), and an Alpha of Arena Commander I, the space sim game within the Star Citizen space sim game (Simception). I don't agree that there is no such thing as a pre-alpha, I've seen a few people espouse the idea that pre-alpha is a new thing. But it's been a thing as long as I remember, and I am no spring chicken.
Does an Alpha need to be QA'ed? I was surprised when I saw this question pop up in RSI General Chat. Why yes, it does. Releasing a barely or nonfunctional application would be useless to anyone, can't effectively test because the known bugs block out finding out anything new and interesting. It would be a waste of effort since staging to servers and manning customer support isn't free.
CryEngine to blame? Well maybe. I am sure CryEngine has it's pros and cons, but it's academic at this point. Swapping engines would delay the game to the point of disaster. Chris Roberts selected CryEngine because of the visual capabilities and the kind of support and partnership he could get from the parent CryTek company. Honestly at the time there was little other in the way of options. Unreal perhaps, but that was between engines and didn't seem to be in the right place at the time. Starting with Unreal 3 and then porting to Unreal 4 by all accounts would have been very time consuming. Unreal is unquestionably a good engine, but honestly even if Star Citizen were starting today, CryEngine might still be the best choice. A few other nascent engines keep popping up as suggestions, but don't think any of them could hold a candle to Cry as far as maturity and support. And don't get me started on how much bigger this project would be if they were making their own engine. Graphical bugs have been wrapped up quickly so far. So in the long run I suspect we'll largely be thankful to CryEngine.
Delays are kind of inevitable, particularily if you aren't padding out your estimates to take account for the inevitable snags. If you read Chris' initial announcement of the 29th date, you'll see the phrase, "if all goes well". How often does that happen?
The DFM was initially fairly strongly estimated to be released in December 2013. But that was a completely different feature set and meant for another purpose. Meant to be a quick and dirty demo that showed off the ships using default Cryengine physics and netcode in order to keep up funding so they could keep the lights on and pay for further progress. With the phenomenal success of the funding, that was no longer needed. Indeed, releasing and supporting that could be a long term drag on development, and an unpolished demo could have the opposite effect on funding and reputation. After polling the community, a delay was made to create what we now know as Arena Commander. Still that estimate was was a few months, which seems almost humorously naive at this point. Next date was Pax East, a reveal on April 10th and release "soon after". Days became weeks became months since Pax East and a date was finally set for May 29th "if all goes well", which it didn't.
So a few days doesn't seem like too much, but it is easy to understand why people are frustrated. What will we see in the future? Something different I expect. Will it be as Ben Lesnick said in Chat recently? "No dates for anything ever ever again." I actually wouldn't mind that, as long as they kept up the communication of progress. Monthly updates for the business, weekly for modules that are nearing release, daily for modules on the cusp of release. Releases done when the product is ready. That wouldn't be so bad.
As I said, this weekend is a critical time to watch, there are concerns that stability might be hard to accomplish with were CIG stands now. They've grown very quickly and been retooling from making demos to actual "production code" it would not be surprising if they had not developed good development standards to deal with unit testing, automated regression testing, change control and the like. These things may be very important to spend effort on in the coming months as it's hard to imagine a huge project like Star Citizen achieving stability without a system like that. I would urge CIG to look at tests suites, Test Driven Development, and Continuous Integration. Release Discipline is a thing as well. There is a worry that shiny new things creep into releases when they are too close to release. I know Chris has a bit of reputation for pushing this and who says no to Chris when he says, yeah put that whizbang in, I know we are in code lockdown but I really want that.
Perhaps the "new guy", Alex Mayberry, will be bringing some of the experience needed for cleaner releases in the future.
UPDATES 2014-06-01: Added disclaimer to top of article, corrected grammar error, fixed misdirected Reddit links and added Reddit discussion of this article.
COMM-LINK - Arena Commander Weekly Report - May 12-16 - (First mention of the 29th date.)