04-06-2003, 03:37 AM
. . . that it was 98% accurate. Until I stopped to think and realized that meant an average of two mistakes per line. ;)
Hi,
I have to give Buzzard a lot of credit, compared to other game companies, for writing code that at least runs without locking up the system or dumping me to the desktop every now and again. So, compare to the industry norm, that makes them excellent. The length of your yardstick determines how tall you are, I guess.
Most of the problems I've seen with Buzzard code is *not* the fault of the individual programmers, it is the lack of programming standards in the company. Things like calculating the values used and displayed in two separate places, making the giving rise to the infamous "lying character screen". Things like placing values in data files, then hard coding the same variables (but with different values) in the code. And so forth. The problem isn't unique to Buzzard, or (unfortunately) even to the game industry. It appears to be simpler, faster, cheaper to go cowboy with the code development. Enforced standards, code reviews, walkthroughs are perceived to be too expensive. OK for the military boys and NASA, but not for non-critical earthbound software. The funny thing is, the companies that have converted to doing it right have learned in every case I've seen reported that discipline actually reduces the overall cost of the projects. Sure, it adds 10% to the coding effort. But that's only one percent to the total effort, and it saves many times that in debugging and revision costs.
But that is all neither here nor there. The fundamental point remains. People are being banned on the basis of information generated by code generated by Buzzard (or Blizzard, if the South branch is involved). The company claim is that no errors are being made because? The code is right? The assumptions are right? The analysis is right? What?
Just what is different in this project from all other Buzzard/Blizzard projects that one is to have vast confidence in the program?
--Pete
Hi,
I have to give Buzzard a lot of credit, compared to other game companies, for writing code that at least runs without locking up the system or dumping me to the desktop every now and again. So, compare to the industry norm, that makes them excellent. The length of your yardstick determines how tall you are, I guess.
Most of the problems I've seen with Buzzard code is *not* the fault of the individual programmers, it is the lack of programming standards in the company. Things like calculating the values used and displayed in two separate places, making the giving rise to the infamous "lying character screen". Things like placing values in data files, then hard coding the same variables (but with different values) in the code. And so forth. The problem isn't unique to Buzzard, or (unfortunately) even to the game industry. It appears to be simpler, faster, cheaper to go cowboy with the code development. Enforced standards, code reviews, walkthroughs are perceived to be too expensive. OK for the military boys and NASA, but not for non-critical earthbound software. The funny thing is, the companies that have converted to doing it right have learned in every case I've seen reported that discipline actually reduces the overall cost of the projects. Sure, it adds 10% to the coding effort. But that's only one percent to the total effort, and it saves many times that in debugging and revision costs.
But that is all neither here nor there. The fundamental point remains. People are being banned on the basis of information generated by code generated by Buzzard (or Blizzard, if the South branch is involved). The company claim is that no errors are being made because? The code is right? The assumptions are right? The analysis is right? What?
Just what is different in this project from all other Buzzard/Blizzard projects that one is to have vast confidence in the program?
--Pete
How big was the aquarium in Noah's ark?