This post was mentioned on Reddit by NippleThief: Pirate away!…

On a totally unrelated note: I have no idea what letters are in the captcha box so here goes at attempt #1.

I agree regarding the ‘trivial’ tests. They are definitely good to have around. A trivial test is better than no test.

My definition of trivial is a “down the middle” tests. I.e. that does the simplest thing possible while also asserting against something meaningful. If anything, they provide a base to work off if nothing else is provided. I used to have to do a lot of picking up of other people’s code (not so much now, thankfully) and writing tests was a lot easier when such tests existed. By extension it was also easier to see how to use their classes.

My problem is when trivial turns to useless. If they don’t prove anything beyond that an exception didn’t occur, the value provided is debatable.

In Pragmatic Unit Testing they suggest you “Spring The Trap” – introduce an error into your codebase and make sure your test picks up on it. This is another technique you can use to make sure your test is doing something meaningful.

I think there is also a distinction between a pointless test and a trivial tests. Trivial tests are good to keep around. They document your learning and understanding of a problem and help you make sure small pieces of the system wire together properly or meet some baseline assumption you have of them. Neal Ford calls them “canary tests” – because they can be the “canary in the coalmine.” Think of running them when you go to upgrade a library to a new version and see if anything breaks.

You make a very good point regarding the tension; the analogy definitely works. Sometimes it does feel a bit unfair.

In particular, it’s very frustrating when you down 3 survivors and come agonisingly close to getting the 4th, but they all get up and find (what feels like) 4 medpacks in the next room. They heal up and carry on like nothing touched them and you have to start again.

I’d liken the interaction with the special infected to boxing — you only need a puncher’s chance to score a knockout. The survivors can be coasting a few rounds up, but in the next moment they lose their concentration, do something dumb and it’s all over. However, when one team is playing the survivors and rushing nicely, totally controlling the game, a disorganised / poorly skilled infected team usually falter. They cannot co-ordinate anything of note.

We once forced a full team rage quit on Blood Harvest 1 after a single survivor round. Ironically, we were trying to make things harder for ourselves (by constantly rushing where it is dumb to do so), but we somehow finished the round in about ~90 seconds (we didn’t stop running). This caused the other team to leave in frustration — they didn’t even get one attack in.

I’ll have to check out Quake Live. I’ll admit that the prospect of playing Q3 again leaves me cold, but it might be worth a go. I always did love RA3 :)

Therefore, the variance in skill doesn’t matter as much as the tension between the two teams. What is really important is the hope, the feeling that it is possible to “score”, even if you don’t. That’s why ragequitters tend to leave on the secound round, because they think like this : “ok they scored, but maybe we will too”. If they don’t, they feel hopeless and leave.

Great matchmaking systems are definitely the next step in multiplayer games, sooner than later it will be unthinkable for a multiplayer game to miss one. It will be of the same importace as a good IA in solo games.

ID has done a terrific job with QuakeLive. The matchmaking system is near perfection for FFA games. All it misses now is an ingame team balancer, a matchmaking system who will also set up the teams for CTF/CA/TDM

I never thought I would play Q3 again but here it is. It has become the perfect game to make quick sessions (10-30 mins). It’s the antithesis of L4D : you know you’re going to play on a quite interesting server in a matter of seconds. It brought Q3 back to life because this particular game is more suited for little sessions (while a L4D game can last more than an hour), therefore you must be able to join and leave good games painlessly. Newbies can play together without being disgusted by demi-god ones, who in turn will have a harder time to dominate everybody since the skills are quite homogeneous.

You never thought so, but in the end it’s really a different game. It’s like maximizing the potential of the original game. I started on low skill servers and my progression to the better ones has been smooth and transparent.

The potential of L4D has a long way to go to be fully exploited. My stats show 47h of gameplay (I thought more), and I’m done with it as it is. My best games were with a french gaming community. We used to play on our own servers, with teamspeak, and more than often the games used to be interesting. But now I’m not willing to waste 15-30 mins without the certitude to find a decent game anymore.

Nice blog btw.

http://johnfmoore.wordpress.com/2009/03/25/unit-testing-for-better-and-for-worse/

Keep up the great writing.

John
http://johnfmoore.wordpress.com