Developer Diary #9

[Mike]

Design

It's been another productive month - the end of pre-production is very much in sight, which means that all of the main systems in the game will be in and working, if not polished to the degree they will be in the final game, and we can set about making the actual content of the game - weapons, levels, story etc.

I'm really pleased with how much fun the game is, even at this stage. Just getting your men to run about and shoot stuff is hugely entertaining, even if I say so myself. 

But the thing that I'm most pleased with so far is the control system. I've worked really hard to keep the control system as simple as possible, while still giving it all the depth and flexibility that it requires to serve the needs of the player, and I think that we're well on the way to achieving that. Games in this genre can easily get submerged under countless keyboard shortcuts and modifiers, so that new players can feel like they have to learn a hundred commands before they can even tell their men to move around. In our game, all of the basic commands, and a large proportion of the advanced commands, can be performed with the mouse alone. Simple acts, such as moving the camera around and ordering men from place to place, or attacking an enemy, are so easy to do they quickly become second nature. The camera being so easy to use has made a big difference to the way I find myself playing the game. I've constantly got the camera on the move, flicking it from place to place, rotating to get the best viewing angle and, most importantly, zooming far up in the air to get an overview of the situation and issue orders before whizzing down to ground level to watch the action unfold from right in the thick of it. You might think of a game's camera control as being peripheral to the game itself, but it really is an intrinsic part of the experience.

Anyway, aside from the control system being ace (which it is, I may have forgotten to mention that), we've made a lot of progress with the AI this month. Lee may go into more detail on this but the short version is that we've spent a lot of time looking at the way that units behave in a given situation. Like most games in this genre, we're giving the player the ability to change this behaviour at any time, via a series of toggles. When I initially designed this part of the game, I suggested a base set of toggles to use, with the intention that as time went on and new situations arose, we'd add to that base set. Which is exactly what has happened; now we've got more than a player could comfortably use in a combat situation so, once we're happy with the way those toggles are working, the next stage is to go through them again, culling any that we don't think are needed any more and combining any that can be combined. We're pretty much at the stage where we've got them all working as we like and we'll be moving onto the culling and combining stage over the next month.

The other main thing I've been doing this month is getting the Design Document back up to date. We've changed a few things, created some other things from scratch and come up with a lot of solutions to long-standing problems, and those all needed writing up; partly so that we can always refer back to the document to find out how we agreed to solve a problem, and partly to make sure that our new systems don't clash with any old ones. When you're coming up with solutions to problems it can be very easy to miss the knock-on effects those solutions can have with existing systems. I find that reading through the document when doing updates can often highlight any issues and give you a chance to fix them before they get too ingrained. Fortunately, even with the big changes we've made this month, there was only one minor issue I found, and with a bit of thought that one was easily fixed.

So, a very successful month all round. If we have a few more of those we'll be finished in no time!

[Lee]

Code

As Mike mentioned, it has been a very good month. We've spent an awful lot of time iterating our systems and getting them to feel right. I now understand why people say that RTS style games are so fiddly to do. There are just so many systems and inter-dependent rules to develop, sets of odd situations that the player and his troops can find themselves in and so on. It has been quite a challenge to keep all of the information in our heads whilst adding the new bits in. I think we made our lives harder by choosing to mix melee units and variable ranged combat in the same game, but the results are definitely worth it. It is great fun to be in a mid-range gun battle with a particularly tough set of opponents and send in a group of guys to attack them from behind with big swords

All of the big systems are now in place, there is one more medium sized system to do (to do with magic), and then we're ready to all sit down and reflect on what we have got. We've talked a lot about what we call pre-production, and to some, it may seem that we have been in this phase for a long time. That's true, but we almost certainly have a different idea of pre-production from most people. We consider production to be the phase when we are adding models, making levels, optimising rendering code and so on. It is specifically *not* working out how the game works, making it fun, answering untold numbers of vital questions and a whole host of other issues some developers seem to leave until late in the day. We have a particular process that we go through when making a game:

Stage 1. Blue-sky
In this stage, we are just trying out ideas to see what might work. When we are happy with a particular direction, we produce a top-line design overview, really just showing the broad brush-strokes of the game.

Stage 2 - Pre-production
This stage takes the design overview and kicks off numerous threads of development: Initial concept and visual style, technology requirements, gameplay prototyping, detailed design, provisional asset lists, full concept art and so on. We then design an example level that contains all (or as many as are practical) gameplay features and begin building that. When that is done, we playtest it and then any ideas or changes feed back into the design process and so on until we have something that we are happy with overall. This is where we are at now. The next step is to enter a slightly more formal iteration loop, and for that, we all sit around a table for a few days, play the demo, go through every feature and mechanic and discuss it, sign it off or agree it needs to be changed. The aim is to start locking things down and finalising the systems. Once all that is done, we produce a final schedule and asset list and enter production. Oh, one thing I should mention is that during this phase, we will do play-testing with people who have never seen it before.

Stage 3 - Production
Here we take the full playable level, full design document, complete schedule and asset list and just get on with it. Due to the nature of this stage, we have a very good idea how long it will take, as we should have removed, or certainly minimised, any uncertainty or dependency. The first step is to produce a skeleton version of the whole game. If the game has 20 levels, then we will make 20 levels and get a rough representation of each one in, work on how the whole thing hangs together and how everything flows. And then it is on with the main meat of the game: making models, animations, textures, shaders, effects, sounds, music, levels, and all that fun stuff.

And then beyond this, we have a testing phase then post-release support and so on.


Anyway, I still have some magic spell code to write so I had better get on with it By this time next month we will have had our big meeting and we will be in full production. Honest...

[Mike]

Replying to a diary post? It's unheard of!

I just had to say that Lee hit the nail on the head with his post; you'd be surprised at how little time is spent on refining the core gameplay aspects of most games.

Maybe it's a symptom of larger teams: you have to keep everyone busy all the time, so you've got people churning out levels, animations, etc. right from the start. That means that it becomes very hard to change any of the initial gameplay decisions when problems arise, or it just becomes apparent that there's a better way to do something; if you do change core features, then most of the levels etc. would have to be redone, and that costs a lot of time and money.

Anyway, whatever the reason, it's very rare to be able to spend the amount of time refining things that we have and I'm very pleased that it appears to be paying off.

[Mal]

ART

Well I have to say that Mike and Lee have given a top appraisal of the past month. Not much I can add to that, perhaps just agreement that it feels great doing things the right way, this is of course a reason we started Games Faction in the first place

I'll keep my diary brief this month as you may have noticed a 'flash' movie of some screen shots and I bet you aren't reading this any more, are you, big nose! If you are reading still, I do apologise.

I've decided to just show these in the forum as a treat for you guys who read here, as it's very much a work in progress, we'll post final screen shots closer to release as our marketing campaign builds.