Frezik
Celtic Anarchy
Anarchy Empire
   |
Posted - 2006.12.29 17:45:00 -
[1]
While I suspect my reply will be a waste of time, I'll do it anyway . . .
You can't just throw new people on a project and expect it to get done faster. Each new person needs to be briefed on the current state of the system, and they also create extra communication paths. Extra communication paths mean more work that doesn't directly translate into higher efficiency. Therefore, there is a point where adding more people actually makes the project later.
Lately, CCP has been tending towards more general systems as replacements for old systems. A good example is contracts. Contracts will eventually be used for a lot more, but for now it's just an escrow replacement. By using it to replace escrow now, they can work out the underlieing bugs before trying to make it do more. Problems are bound to come up, but doing it this way means you get hit with fewer bugs at once.
You can't get a better testing environment than the >100k users of TQ. You can do simulations, do open testing on singularity, and a bunch of other things to get the number of bugs down, but ultimately you can't simulate thousands of people pounding on it all at once in the real game.
Software, as a discipline, is not comparable to other types of engineering yet. An engineering on the scale of, say, the Golden Gate Bridge is a massive undertaking, but humanity has been building bridges for a long time, and we can build on previous experience. Computer software engineering is roughly 50 years old, and some lessons taken from general engineering (like the Waterfall Development Model) have proven to be failures when applied to software in practice. The immaturity of the discipline is a major problem, not just for CCP, but across the entire industry. But people won't wait for the discipline to mature--they want their projects done now. So there's little choice but to accept the problem and try to do the best that can be done.
In short, I think CCP's development practices are the best that can be hoped for given the state of the entire software industry.
|