AvP2 Team Fortress

Aliens vs. Predator 2 mod |

So we want to bring Team Fortress to AvP2. It shouldn't be too difficult, all of the weapons we need are already in AvP2, and configuring armor and run speeds and whatnot isn't too hard. What we need, however, is input on how this should go down, and modders to make the more difficult things happen, like possibly adding new weapons and more importantly, new GAME MODES, like Capture the Flag or Assault, things that teams do that aren't Team Deathmatch. However, this is AvP2 Fortress, so everything should have an AvP2 twist to it. Like Preds and Aliens on the teams realistically, but not whorishly. No Predator rocket-launcher guys, for example, but a Predator as the Spy, a class that's supposed to cloak and primarily use a knife anyway. -- SMID

Post feature Report RSS AvP2 Fortress: post mortem

The title itself shouldn't be viewed as the funeral of AvP2 Fortress. Instead, a post mortem for a mod usually contains the thoughts of the developers of what went good and what could be done better for that mod.

Posted by on

This may seem a bit weird, since usually post mortems are written for mods that got canceled, at one point in their development. Although that was not the case for AvP2 Fortress, still, there are some valuable lessons we learned from the development period. The pre-production stage. Pre-production started in november 2007. I was still working in Ballistic 2; one feature from that mod (the Javelin weapon, first fetured in UHM) had a lot of potential to fill in the role for the feature of dropping sentries, ammo and so on. The idea of such a mod was received with a healthy dose of skepticism; I mean we were planning such a mod, six years after the game was released: The question is: Woulden't it be the best to "kill your darling" and plan this mod for the upcoming Aliens- game instead? I know that's what I would do. But we were both lucky, and unfortunate. Aliens - Colonial Marines will be released only in Q1: 2009. Lucky because we were right to stick with AvP2, unfortunate because, well, Aliens - Colonial Marines was delayed to 2009. So, the only ammount of time I had was actually 'leftovers' from Ballistic 2. Still, being strong influenced by the work environment from my current job, I decided to use that very little time to write down some specifications for the mod:
Lesson Learned: Don't even think about starting a big project without a plan. The "let's ride this bull and see where it throws us off" works only for small projects (or for big projects, if you're in for just a very brief "bull ride"). The "If you fail to plan, you plan to fail" saying was not intended only for C&C Generals: Zero Hour.
In the spring of 2008, after 'recovering' from Ballistic 2 and having a decent document as guidance, the javelin like feature was evolved in the CPandora projectile class, the backbone for the special features of Fortress. It was mostly a fun period, because it was all about experimentation.
The alpha stage The days before going alpha got pretty tough. Only then I realised that the scale of this mod is big. Too big to be handled by one person. I mean at that point I wanted to make a playable map, something better than the cube I was using to test various features:
The old 'A Similar Fate' map, used heavily in the alpha stage; currently showing one of the paths connecting the two bases.
'Our Collective Fate' was the natural step forward from 'A Similar Fate'. It features more detailed geometry, and the neutral flag is easier to access.
'A similar fate' was the experimental map we used. It featured two 'A lesser fate' maps connected together by three paths. It was a crude map, and the fact that it remained virtually unchanged during the alpha stage, now points out even more that Fortress is bigger than one person.
Lesson Learned: Don't be affraid to ask for help. Large projects need larger teams, to share the work load. Even if you can make decent maps, and you have a natural talent for 2d artwork and skins, and on top of it you're an exceptional coder, if you plan to release your project within a decade from it's announcement, you will need help. Not that I want to offend anybody, but "Rambo" is just a movie.
But in this period, Windebieste announced that he considered AvP2 Fortress as his next project. Having the mapping effort of my shoulders meant a lot to me, and also 'instructed' by the podcasts at moddb.com, I started to work on the mod's site. Although public relations also took time, the alpha period provided a lot of good times. It's when the features of the mod got posted on youtube.com for the very first time. The mod got more and more publicity.
Lesson Learned:Do "smart" advertising. Promote all the features that are already completed, or that are functionally complete but you need only game resources to replace the current placeholders. This way you'll avoid embarassing situations, where features that were advertised don't make it to the final release. Another reason to advertise is to attract the attention of other modders. It's a good idea to use the alpha version of the mod for recruiting.
Another good thing that appeared during the alpha stage was the automated build system. It consisted in a collection of .bat files and the end result was the installer the mod currently has. The installer itself was then perfected durring the beta testing phase.
Lesson Learned:Human beings make mistakes. Try to automate the packaging of the mod, because it will save you hours of double-checking if all the resources are in. It's a "fire and forget" type of tool, once you have it set up right, you have one less thing to worry about. Also, if possible, deliver installer packages. The regular user is much happier to just double click and go "Next->Next->Next" than to extraxt an archive, add some parameters to the command line and cross his fingers in the hope it'll work.
The beta testing period By far, the toughest period. Besides coding and PR, arranging for the tests themselves took its share of time. Getting people in the beta program was a stressing task too; I registered to several active AvP2 communities: - I started by writting down the names of the clans I saw when playing online - next step was to google the clans to see if they had their own forums. - yep, next step was to join the most active forums found at step 2, and launch the spam: AvP2 Fortress is looking for beta testers!
Lesson Learned:Just like the second lesson, if you don't have enough time, try to recruit a person to handle the beta testing campaign. Everything from searching for testers to setting up the server to collecting the bugs from the testers, it takes time. So make sure you have one person dedicated to handle and organize the beta testing campaign.
The first person to join the beta program however did not come as a result of the posts on those forums. Vemarkis noticed the mod on moddb.com, and he was the first (and for the first tests also the only) tester to join the team. More testers joined afterwards, but these first tests, when only me, Windebieste and Vemarkis did online testing did point out that something I did was wrong. The recruiting strategy was not that good. Perhaps I should have started first with the communities I was already active on, and individually contact people from planetavp.com, or lthfaq.com. I honestly don't know, and your input will definately help. Also during the beta period, I made the mistake that is almost my trademark: I got overconfident. The result? Beta 5 speaks for itself: plenty of crashes, people getting stuck in spawn after pressing F1. It was horrible. I was able to fix the crashes the next day (heck, I knew what I have altered, and reviewing the code pointed out a mistake I did), so beta 5 had a lifespan of one day. But it was enough to make me be more careful.

Lesson
Learned: Be careful! It's very easy to have a critical mistake slip through, so don't get over confident. If the mod is single player, then play the entire campaign. If the mod is multiplayer, make sure you do multiplayer tests with enough players, since some bugs cannot be detected with only few players in the map.

Overall, the mod ended up getting me more and more tired. As a result, for about 4 weeks, no news were published, and the beta versions had only minor fixes. Unfortunately, PR got neglected. Still, one thing that went good (besides the mod slowly getting polished) was the fact that we prepared an SDK. At Windebieste's suggestion, we prepared a collection of files that would make mapping for AvP2 Fortress easier. Do not neglect PR. The only way to have the mod spreaded in the community is by keeping an open communication channel with it. Another thing you can do is have the community involved in creating contents for the mod. If you can, build some sort of SDK for your mod, it may give it longer legs.

Lesson
Learned: As a conclusion, what I feel we did good: - use the automatic build system: it made packaging the mod so much easier. - intensive PR during the alpha stage - used a versioning control system (it saved my ass more than once) - we split the tasks. That means Windebieste doing all te maps (a pretty big task if you ask me, so, thank you), and accepting resources from the community. The flag is a great example as to how several people can work on one item and make it just perfect. - we mailed each other. It means a lot to always know how the mod is doing, and communication plays an important part in this. We could have done better: - recruiting beta testers. Perhaps a more open approach would get better results. - PR during beta. It's a simple rule: if there are no more news about the mod, it gets forgotten. - announce beta tests in time. A known schedule is good, but sending mails one day before is even better. People may forget the schedule, and in such cases a mail notification is the better solution.

But the most important lesson when making a mod:

Have fun doing it!

-- Herr_Alien
Post a comment

Your comment will be anonymous unless you join the community. Or sign in with your social account: