Episódio 6

Monte: Sure, thanks Danielle. Well I’m a lifelong video game player. I started way back as early as I can remember and started on the Atari and had Atari computers and learned to program pretty early on. So I went through a couple of different majors in college and eventually popped out with a degree in computer engineering and with my shiny new degree went out and took a really lousy job. It got better though, and after that I was working on product development writing software. And I enjoyed that very much, but my kinda life’s passion, what I went home and did was play video games. I still, you know, all through college, all through high school, earlier, even after I got out what really drove me was games.

So after a couple of years of writing software, I thought, you know, I can do this game thing and I was living in Chicago at the time, I’m a native Chicagoan, and I found a little teen tiny game company out there called Game Refuge, and the founders of that were old arcade veterans. They had done some really classic arcade games: they worked on Rampage, and Tron, and Spy Hunter, some cool old stuff like that. Kind of a fun group to work with and they took a chance on me it was kind of cool, so I spent a couple of years there. They were a small shop though and after awhile we were looking, it was unclear if we were going to have some future projects. So I had to look a little bit bigger and better, and by that time I had already killed more hours than I care to talk about playing Diablo and Starcraft. So when the time came to start looking, I thought you know, I really love those Blizzard games, let me see if there is something I can offer Blizzard. And, there was a job for a Tools Programmer, at the time, and what I was doing at the other place was writing a lot of tools and it was arcade development so it’s kind of different from what we do here but a lot of the skills lined up.

So I applied and I got a job as a tools programmer and at the time, this was in 2000, I was working on technology that was shared between World of Warcraft, which was a tiny budding little project at the time, and Warcraft 3 which was in full development. So, for about a year we worked on shared technology and then we split the techs, and I finished up on War 3 and worked on Warcraft 3 and its expansion, and then Starcraft 2, and now I work on World of Warcraft.

Monte: Sure well we’re pretty close to the end of Lich King (Wrath of the Lich King) now, now that it’s in beta, so our days are kind of anything but normal right now. So we have kind of a mix. We have a lot of tools and WoW has been in development for a long time so there’s stuff that’s older, stuff that’s newer, and there’s a lot of what we call fire fights. Sometimes there will be a bug that prevents someone from working and we’ll have to drop something and take a look at that right away. Especially if someone thinks they are losing data, that’s very serious to us, so we want to make sure that their work flows as smooth as possible. If we make the work flow good, the designers make great stuff with WoW. So it’s really important for us to try to optimize that, make that really good.

So yesterday we had a bunch of fire fights. In between those fire fights we’re looking at bugs and getting bug reports from the field. We’re also trying to work strategically. Kind of, for the lifetime of WoW, what are the cool things we want to do? What would make everyone’s work faster? What would make everyone’s work better? Our goal is to make the tools fun and easy to use and we want the designers to come into work and want to use our stuff so we spend a lot of time thinking strategically about how to make the UI better, about how to make parts of the program faster, about how to give them better feedback about what is going on, let them see encounters earlier. Let them work with more textures. Let them work with more creatures. Just all of these types of things that we look at to try to make it as cool as possible for them to build cool stuff for everyone to play.

Monte: Oh, those are great questions. And we agonize over that frequently to be honest. The types of things we haven’t produced in house, we haven’t produced anything like Photoshop or anything like 3D-Studio Max or Maya. Both of those, or sorry, all three of those products have a really long development history and people much smarter than me have spent lots and lots and lots of time working on them. And frankly, they’re so good at what they do, I don’t think it would make sense for us to try to re-implement that in house, plus in those cases, those products are very flexible. We’re able to add on to them as necessary in able to get data from them. It’s not always puppies and kittens, we sometimes have some problems, but for the most part, it really makes sense to leverage that big body of work.

Now the flip side of that, why don’t we just use all external tools is, well… Part of the Blizzard character is, there’s a certain flavor to what we do. Our core competency is building real expressive art and animation and stuff that really captures the Blizzard character. And to do that we need certain levels of flexibility. That means, it’s very strange, you would really need to understand the inner workings of what we do, but in order to capture a lot of that performance that the animators do, that our artists do, we really want to be able to really really have fine control over the data that comes from these programs. So we end up writing our own tools in order to capture things that are maybe just handled too generically in some of these other tools.

Monte: Yeah, absolutely. Yeah, very much so. There’s a lot of that, plus you know, we have a work flow that’s probably different … It’s related to other work flows in the gaming industry but it’s very very specific to how we build WoW. So we want to build customized tools that help our people who are most familiar with their work flow and how they want to work, we can really maximize the productivity and help them make cool stuff faster. That’s ultimately what my goal is as the lead of this group, is to help them just make the coolest stuff they can as quickly as they can and we can do that by really fine tuning, customizing the tools to how they make the stuff.

Monte: My most challenging task at work, I would say really taking over the tools team and just being in charge of WoW Edit. WoW Edit is, it’s a product and it’s a product that we ship before WoW ships, we have to ship it beforehand and our customers live on our floor with us. You know they don’t write on the forums, they come and line up outside my office if there’s...

Monte: They sure can and they sure do. So I think the toughest project we have is ongoing and that is to make WoW Edit really a world class, truly fun, truly productive development environment for all of the people who need to work on WoW.

Monte: So that’s a great question, Danielle, and I would absolutely recommend Computer Science. I really think that getting a degree in Computer Science really really gives a good strong foundation for the kind of skills you need to come work on my team. Within that I do have a couple specific recommendations. We use lots of C++, I get the question all the time, what kind of languages should I learn? I’m working with C Sharp. I’m working Java. I’m working with C++. I look for generalists, you know, people who are good on their feet , who can kind of use the right tool to solve a problem that we have. By far the largest set of that happens in C++, so I would recommend, learn C++. The other thing I’d recommend is learn math. Math turns out to be really important : linear algebra, trigonometry, a lot of stuff we checked out in turns out to kinda come back in force.

Monte: Oh, it really does. Especially when we do 3D graphics. It’s really really important. Trig is important. Linear Algebra. Matrix Math. That stuff kind of comes back in force. It sneaks up on you too, you think “I won’t need this to program” but then all of a sudden, “hey!”

Monte: Yeah finding the math book is like the moldy tome out on the shelf there. Crack it open, hopefully the pages are still intact and see if you can find the uh...

Monte: yeah

Monte: Yeah right, the chart in the little overleaf, look for that.

But, the math turns out to be really important, so I would definitely tell people to make sure to study the math.

So, a couple of more pieces of advice. Since I look for generalists or people who are kind of quick on their feet to solve problems, there are a couple of other languages that can be sort of useful doing one off development or data processing kind of stuff, so I look for people who have some background in something else maybe, maybe a little bit of Perl or Python or Ruby, those all turn out… Those are popular today and they turn out to be really useful for solving problems. So I think that’s some good stuff, I look for that.

And my last piece of advice for getting involved is to get involved. I know that sounds kind of like a totology there but there’s a lot of opportunities to kind of, without having to go out and actually get a job, there’s a lot of opportunities to get involved in game development. For instance a lot of people who want to maybe get involved with WoW, we recommend that you… Like if you’re not sure, you’ve never worked in programming and you’re not sure if you’ll like it, not sure if you do, you can try writing an add on for instance for WoW. And the UI part is written in XML and the programming part is written in LUA, which is another programming language that’s pretty easy to pick up. It’s similar to some of the other ones we work with and it’s a great way to kind of get a feel for the kind of work we do. And when we interview, we look for those sorts of sets of skills and some interest. So if you’ve written a WoW add on, or a War 3 modification, or like a conversion for Half-Life or a Quake mod or those kinds of things, those go a long way towards helping get into our industry.

Dustin Browder: It’s a pretty organic concept. I think it sort of depends, right Sammy, on which way we go. Sometimes we start with the art first, sometimes we move from the design first and sometimes we have different, you know, we have a piece of art we have a piece of design and we try to marry them together.

Sam Didier: Yeah a lot of the times the artists will just have a cool idea we’ll show it around and everyone will go “Yeah we got to get that in the game.” And then sometimes the designers will have an idea or two “Hey we need something that will ‘this’ and ‘this’ and then the art team will start concepting things up and from there we just let it roll and iterate, iterate, and iterate and at the end of the day we have a finished unit that we will redo three or four times.

Dustin Browder: (laughs) Right and be balanced a thousand times before it’s finally done.

Sam Didier: Well we had an idea, on the Terran side, we had the old Goliath. It was cool. We had the Valkerie. It was cool. But we wanted to kind of make the Terrans a little bit more new. We wanted to add some more transforming to the Terran side so we said “Okay well what if each of those was one stage of the unit and then it transforms?” So now you have an air attacker, you have a ground attacker, whatever they end up doing at the end of the day whether it’s machine guns on the ground, missiles in the air, whatever, we knew we wanted to have some kind of transforming robot. That was one of the areas of sci-fi lore that we hadn’t had. We all grew up in the eighties and nineties watching transformers so that was something really dear to our heart that we wanted to put in.

Sam Didier: STARSCREAM!

Dustin Browder: (laughs) Yeah I think units like the mothership, I know when I first got to the studio from the original movie for the beginning of Starcraft had a very huge Protoss ship in it that shot a terrible beam and destroyed the small terrain salvage ship and this was a unit that was part of the lore but we never actually saw in the game so the idea of a mothership, I think, had been floating around the studio for quite some time just based on that opening cinematic on that really cool piece of art. And so there was an example of a unit that sort of came from art and lore first and then we’d been working on different mechanics for it now for quite some time to make it work in the game.

Dustin Browder: It’s usually pretty obvious to me when we got something good. We got a nice piece of art really if we don’t have mechanics for it you’re pretty enthusiastic to come up with cool mechanics for a nice piece of art. At the same time, because we’ve been able to play our multiplayer for so long if we’ve got a nice game mechanic, I think a lot of the art guys are pretty enthusiastic to come up with cool art to match that game mechanic. So when you’ve got something cool it’s pretty easy to get enthusiastic about it. It’s just making sure we have enough of those in the game to make it exciting.

Dustin Browder: So the Thor does two things for the Terrans: It gives them a sort of tip of the sword kind of unit, something you can push up front that can take a lot of damage from enemy fire. It’s very survivable, very tough, very hard to kill. And you can use that sort of push past an artillery barrage or push directly into an enemy base to sort of lead your smaller lighter marines into battle. And it’s also got these huge cannons on its back that can strike against air targets so you can use the Thor to defend your forces from enemy air threats and it’s a very powerful weapon in both of these roles. So it really gives the Terrans some additional flexibility in the battlefield they haven’t really had before.

Sam Didier: Yeah basically, I think I covered a little bit early, but we’ll have a bunch of our artists do some concepts and we’ll look at the different pictures and say ‘Oh I like that. Oh that part looks cool. We should add that to this one over here.’ So after a couple of iterations on the concept we’ll settle on one and then we’ll sic one of our artist to start modeling it. And they’ll model it up, they’ll start mapping it, we’ll have our texture guy, if it’s not the same guy who modeled it, start working on it and then it’s ready to be handed off to the animators. And the animators will hit up and we’ll go through various stages of the animation, we’ll have a, here’s a quick walk skittering across, here’s one that makes it look like it’s lumbering. We’ll see which one looks the best because the walk is basically what you see the most with all these units. Then after that we’ll pick one and then we’ll go on and settle on an attack. ‘Oh we want it to really reach out far.’ We want it to ‘this’ or ‘that.’ After that the unit’s all done then we’ll move in on to the portrait. This is where you get all the character on the unit. You know you can’t really see their faces on the field of battle so we have the little portraits there. And that’s where we’ll come up with, ‘Okay the roach needs to have this. We want to focus more on little beady eyes.’ And this and that and lots of little teeth. And so after that, we may need to go back to the unit and tweak the face that’s on the unit to match more to the portrait. But after that it’s pretty much done and then, barring any design changes or if we’re allowed to add more polygons, we hopefully have a finished unit there and then we’re done with them.

Sam Didier: About six years.

Dustin Browder: (Laughs) It really does depend. Sometimes it does take quite a long time. A unit like the baneling for instance went through the whole process extremely quickly and hasn’t had any significant changes since it first went into the game. More complicated units like a Mothership or a Viking or a Reaper sometimes go through some significant changes as they go throughout the process and they can take a lot longer especially if we end up moving around the design a little bit. If we end up changing some of the core mechanics on the unit like we did with the roach where we changed it from being a melee attacker to be a ranged attacker that obviously has an impact on the art and it obviously lengthens the amount of time it takes them to make that unit work in the game.

Sam Didier: Yeah we have a lot, even now, in games that people are playing now we have a lot of placeholder art still. And we have that in there so that if the designers have a new idea they want to try, we’ll come up with some place holder art that looks cool but it’s not one hundred percent polished because there is no reason to polish this unit all the way to one hundred percent if we’re going to end up taking it out or...

Dustin Browder: If we’re going to chuck it or change it or...

Sam Didier: Yeah, we’re still going over, and I’ll talk with Dustin say,

‘How are we on this now? Are we good to go? ‘

‘We’re good to go.‘

‘Okay. Ninety percent or a hundred? ‘

‘Ninety percent.’

‘Okay Ninety percent is good enough we’re going to finish it off.’

Sam Didier: Well if you want something that’s totally bean ‘countery’, you know, predictable and kind of boring then yeah the design one is the best.

Dustin Browder: And if you want something that doesn’t really play and it’s just pretty but it’s no fun then I think you want to go with the art way.

Sam Didier: : Oh is that how it’s going to be now?

Dustin Browder: That’s how it’s going to be.

Sam Didier: I don’t need a weapon for him! Please.

Sam Didier: All right, thank you. Dustin. I’ll be talking to you later.

Anthony: It’s going really good. I got my teeth drilled today but the Novocain has worn off, and I’m ready to go.

Anthony: Actually no Gravity Incorporated was the first game company, if you could call it that, that I worked at. That was a little start up company in San Francisco, and that was around ’94-’95. They were making a real time 3D game called Banzai Bug, which we actually put out – hit the shelves and everything. Probably sold 5,000 copies but, you know, it was a good experience.

Anthony: Yeah, let me see let me look at my notes. The first thing I worked on I remember asking my boss, “Hey what can I do?”, after being there a few days getting situated, and that was Erich Schaefer at the time. He and Stieg [Hedlund] said they needed some type of flying insect monster for Act 3, so I came up with the mosquito demon. What’s now called the mosquito demon, and that was the very first thing I worked on. It was kind of for a swampy area so I figured that was appropriate, a mosquitoish-type monster, and yeah that was the very first thing I worked on.

Anthony: Probably the most notable monster I worked on was Mephisto. Phil Shenk had concepted Mephisto, so I already had something to start with. I modeled and textured him and set up the initial rigging, and then handed that off to Cheeming Boey to animate. I actually helped a little bit with the effects as well, the effects that are at the base of Mephisto. He has like swirling smokey skulls that kind of swirl around the bottom of him as he’s moving around.

Anthony: Yeah things back then were quite different. Most of the time the artist did everything from concept, to getting it you know with effects and everything in the game. So you did your concept stage, your modeling, your texturing, your rigging, and your animation and then effects. You know whatever skill effects you may have needed for the monster or for the character, and then got that in to the game. On some characters we decided ok let’s have one guy animate it, and I’ll model and texture it, and we started to do a little bit of that back then. It was very different from what we do now.

Anthony: Yeah now ... you know modern art production [and] pipelines for games are way more specialized. They have to be, it’s just the nature of getting things done. So definitely on our team we have things more broken down, not so much as some companies I hear about, but you know I do modeling and texturing mostly. I do some concept work. I help with effects sometimes if the guys need certain things modeled or textured for their effects. The animators pretty much animate all the time, but they also contribute concepts and stuff like that as well. Like sometimes we just want to brainstorm on new ideas, you know everyone will pitch in on the concept phase. But for the most part things are definitely more specialized; you know we have a team that’s just dedicated to rigging.

Anthony: Setting up the characters with skeletons and enveloping the characters, and stuff like that. Making sure things are named properly so they export correctly, and all that kind of stuff.

Anthony: Well the creation process was fairly different, like I said before, a lot of the artists did everything from the ground up. Then there was the sprite rendering part of it, you would have to pre-render your assets, and so basically you’d wait forever for all the stuff to render out, and with the heroes you’d have to render out all of their body parts separately. Things like the gloves and the boots, things that would change on the character. Then you would have to composite those together, for every direction he was rendered in for every frame of animation. So if he’s rendered in sixteen different directions and he has a … whatever, an eight frame animation, well you have to go through all that and composite all those separate elements in the right order for it to look correct on the screen. So you don’t have an arm that’s supposed to be behind the body in front of the body. And… that was a real tedious process.

So thankfully with real time 3D we don’t have to worry about that, and that’s kind of what I was getting at, with things being a lot easier. Also your animation frame count is not an issue any more. We can have nice fluid, smooth animations now whereas before they were really choppy because, you know, we were only given six or seven frames for a walk cycle. You need a lot more than that to make a convincing walk cycle, but you had to make it look its best within the limitations. So the technical part of it was very different, the creation part of it… I think back then things were a bit more tight-knit and collaborative I guess. You know because the team was smaller, you know when you came up with ideas or designs, there was just a lot more back and forth with designers and artists and programmers. It really felt like everybody was more cohesive in that sense, whereas now the department is bigger, there’s more people on the team, it’s more specialized. There still is that collaboration, I mean I can get up and go talk to anybody on the team that I want to, but when they’re way down the hall you start to get lazy and you don’t want to go chit-chat with somebody. You want to get your tasks done. The design process is a little different, but we still, when we come up with an asset now, there’s a particular monster needed for an act for instance, before we really start working on the asset we’ll get together with a designer, a technical artist, character artist, and producer, anybody that needs to be involved and we’ll talk about it and spin ideas around and make sure we can troubleshoot anything we can in the beginning stages. Come up with ideas for how it may die, what else can it do that would be really cool, and we didn’t have that kind of organized meeting when we were working on Diablo II or the expansion pack.

Anthony: It was more... I don’t want to say disorganized, but it was more freeform back then. It was kind of just a natural riffing off each other, and not an official kind of “Hey let’s get together and make sure we’ve discussed everything that needs to be discussed” and sign off on that and move forward. It was just more organic.

Anthony: Yeah definitely, but there was a real charm that came with that too.

Anthony: The female witch doctor’s heavy armor set, her look is my design and work.

Anthony: Yes, yes. And the scavenger monster.

Anthony: Yeah the little burrowing guys. The Corpulent, the guy that blows himself up, I did some retexturing work on that. Let’s see, the tree that explodes, I helped with some effects work on that. I helped provide some modeling/texturing needs for the effects team. The shaman goatman I built and textured, the other goatmen I did some touch up on, their textures. The ghosts, I created and modeled and textured the ghosts.

Anthony: Yeah exactly. I think that’s all I can think of right now.

Anthony: Yes I did. I modeled and textured those off of Josh Tallman’s initial concept design.

Alex Afrasiabi: So Turalyon, last I heard, he was locked away in some legion home world but you know that’s just a rumor. But really there are plans for Turalyon, it’s one of those things that we keep every content patch we try and work out something. Invariably it goes another way and so we’re hopeful that eventually we’ll get to him.

Alex Afrasiabi: I personally like the new model much better than the Night Elf that we had but we are getting a new custom built model which looks amazing.

Alex Afrasiabi: I think it’s gone really well. If you start as a Death Knight, the first person you see is the Lich King. Going through our zones, Howling Fjord, Borean Tundra, Dragonblight, Grizzly [Hills], Zul’drak, I think almost every zone has at least one encounter with the Lich King, some with Prince Arthas himself - the Alliance get a series in Dragonblight for that - so I think it’s gone really well. Then in Icecrown again, you’ll get several interactions with the Lich King and then going forward in patch 3.1 and beyond there will just be more.

Alex Afrasiabi: Thank you.

Dustin Browder: : I’m not really sure what an end game unit is, I guess it would be like a battlecruiser or a carrier or a thor or ultralisk, I guess it’s what we’re talking about here. But the hydralisk has always been anti-armor, even in the original Starcraft had a strong anti-armor role.

The marauder of course gives the Terrans some additional anti-armor capability from the barracks which we felt was very important since we moved the ghost back so far allowing the ghost to fulfill more of a core role on the battlefield for the Terrans. We didn’t think that the ghost made a lot of sense as fundamentally anti-armor with his sniper rifle. If he’s going to be anti biological and sort of taking head shots and taking out individual targets then we really needed the Terrans to have some way to deal with stalkers and immortals and to a lesser extent siege tanks and roaches. So that’s really what the marauder is pointed at, the marauder is pointed more at stalkers and roaches and you’ll see marauders used in that role. They also can tear apart a colossus or thor which is an added advantage for the Terrans from the barracks again.

One of our ideas initially for the Terrans for the infantry for the Terrans was to make sure that the infantry were viable throughout the whole game whereas much as they possibly could. Obviously at a certain point when a lot of psi-storm comes into play or there’s a lot of splash from enemy weapons fire you’re going to see the infantry start coming off of the map very quickly but we really wanted to make sure that the infantry were viable throughout the gameplay experience and that’s one of the things that the marauder is getting for, he’ tough, he can take a lot of hits, and he dish out a lot of damage to some heavily armored units and so if the threat is like roaches, and immortals, and stalkers the marauder has a lot of firepower.

And the hydralisk has a similar kind of role in that sense, obviously the hydra can shoot air, he is also a lot softer than a marauder but the hydra has some very powerful abilities to deal with some of these heavily armored threats that are very low down in the tech tree. Units like the thor, the mothership and the colossus all have their roles and we see them in games probably about a half or a third of the games that we see will actually see these units even come into play. And when they come into play they can be extremely dangerous of course but there’s a lot of counters already in play and the marauder is just one of the choices you have when dealing with an enemy unit like the thor. You can also obviously bring in mass siege tanks, even mass marines ultimately can be very effective against a small groups of thors because the thor is just not doing enough splash, he can absolutely obliterate a marine in a couple shots but there are simply too many marines at that point in the game sometimes for a thor to tear them all down so we’re really not worried at this point of these end-game units sort of eclipsing the core gameplay experience. They’re definitely finding their place, they definitely have their uses but they are definitely something you can counter and the marauder is part of that counter list but it’s not really the reason the marauder exists. The marauder really exists for stalkers, immortals, and roaches. Threats like this that are massable, can be fast but they’re heavily armored, they are something that a marine has trouble dealing with.

Dustin Browder: Alright, thank you.

Jay Wilson: Well what we want to do with that, especially for particular bosses is have when essentially when the player is very low on health and they’re about to die the boss essentially checks like when it attacks you, did I just do enough damage to kill you? And if so then instead of just doing his normal attack he’d actually play some kind of special I pick you up and eat you or I throw you up in the air and knock you around like a baseball or something like that. It’s a system that really our announcement video we tested it for the first time, mainly just to see – can we actually do this kind of animation interaction between the characters but the actual system itself is still not in there but we do plan to do that and that’s primarily where. We might do some other things, we’ve talked about physics based deaths we’ve talked about having the characters if they like die to a cold monster he might completely freeze solid into a statue and then shatter or things like that but we haven’t decided at this point if that’s exactly what we’re going to do.