| From the
keyboard of Corey Cole, System Software Engineer February 16-20, 1998 Hi, I’m Corey Cole, mild-mannered game programmer. Sometimes I’m a vicious game designer (co-designer on Shannara and QG 1, 2, and 4, plus some up-front design work on 3 and 5, designer of Castle of Dr. Brain), but I’m mostly wearing my programmer hat these days. (That’s the black hat with a bunch of Yosemite and Saturn pins in the band.) My first response to being asked for a diary for the "week" of 2/16 – 2/20 was, "What, they think a week is only 5 days long? Not for a game developer, it isn’t!" But that’s all right; we wouldn’t want to discourage any of you aspiring game-makers. Monday, February 16, 1998 Lori and I arrive at our usual 8 a.m. starting time. We do things a little differently in the mountains – in San Jose, programmers drift in around 10 a.m. on a good day. Actually, we aren’t all that virtuous. We have an 11-year-old son, Michael, who needs to get to school at 8. Since the school busses can’t handle our dirt road and the school is right behind Yosemite Entertainment, we get in early most days. El Niņo has been dumping rain on us for several weeks, so we get to play grand slalom around the ruts and potholes twice a day. Shannon Haynes drops in to get an update on my schedule. Each Monday I tell her that my "template room" is nearly finished and ought to be complete in a day or two. Unfortunately, one of the most important skills for a programmer is a good memory. As a result, I remember how many Mondays I’ve said the same thing and feel guilty. It’s amazing how much of the infrastructure has to be developed before the simplest game operation can be specified. One week we needed "scripts" for sequential actions, then it turned out some art needed to be tweaked. This week I need some new hooks into the Inventory Manager. I spend the rest of the day getting the room and the Inventory Manager to talk to each other. Tuesday, February 17, 1998 I talk over some systems programming issues with Eric Lengyel, our graphics whiz. Thanks to Eric, Quest for Glory V will release simultaneously on Windows 95 and Macintosh. Back in November Eric told me that the only problem with this game was that it was a two year project with a one year schedule. This sounded about right; welcome to the game biz. The good news since then is that Sierra has seen what we are accomplishing with Dragon Fire and has added the time we need to do it right. C++ is significantly more powerful than SCI (Sierra’s in-house development language which we used for Quest for Glory 1-4), but the power comes at a price. We "lost" the years of development work that had tuned SCI into an ideal engine for writing graphics adventure games. Of course, Quest for Glory V: Dragon Fire is much more than an adventure game; we’re now starting to reap the benefits of C++ power. I manage to get in a few more lines of room code. Tracing a mysterious crash leads me to discover a missing art resource – a "talking head" hasn’t been fully animated yet. We throw in a quick fix to keep this from crashing the game, then I go on to the next task. Wednesday, February 18, 1998 OK, right back to it. Now I discover that the clever way we put character dialog and room object messages in separate files might not have been quite as clever as we thought. I can’t get a town guard to say something in response to the player putting the hero to bed for the night. A quick talk with Bob Heitman (one of the original SCI authors, now back working with Sierra as Lori and I are) and Frank Kane (the message editor guru) results in a request for a sweeping change in how messages are named. Frank kicks out the change in no time flat, then I spend twice as long letting all the other game programmers know how to change their code to make it work again. Lori tells me to make sure I’m home at a reasonable hour. No problem. Midnight is a reasonable hour for a programmer, isn’t it? Thursday, February 19, 1998 The new message system really helps. I get a whole lot more of the bedroom programmed, then have an application programmer meeting to discuss the new features and any problems people are having. Everybody seems to be making good progress. Selva, a contractor from India, has managed to get a chest to explode with an impressive sparkle effect. I immediately steal the code for something else. (Did I mention that another important attribute of a good programmer – or a designer – is the ability to steal from the best?) Lori needs to know how to write up a particularly complex set of interactions so a programmer can make it work. What happens if you knock out one guard, but ignore the other one, and maybe hide the unconscious guard and maybe not? Should a hero be able to slit the guard’s throat? (I say "No way!") We spend a couple of hours discussing the setup, then I help her write some pseudo-code. Friday, February 20, 1998 Code like mad. This time I discover that my code for the player using inventory objects wasn’t notifying the room which object was used. I slip that in between half a dozen short meetings helping the other staff programmers get their rooms working. They don’t all have to be 60-hour weeks. We head out in the early evening
for a weekend of AD&D and Magic: the Gathering at a San Francisco area
convention. I scribble a bunch of Post-It notes before I leave so I’ll have an idea
of what to work on next when I get back to it next week. |
|
Next >> |
|
Interviews | Design | Heroes | Macintosh | Timeline
|