The Noob Perspective: iterative RPG design

What could be more noob than to purchase an indy game engine license in order to, supposedly, develop an RPG with no experience beyond modding? Yes, folks, you have here before you the prototypical noob. Plenty of dreams and wacky ideas, but little grasp of the scope of the task at hand. And worse, no design and obviously no team.

I have been a dedicated single player RPG gamer since I got my first boxed set of D&D (pre AD&D) as a tike only to learn that almost nobody else in the neighborhood had ever heard of it and those that had thought it was a cult (like, Mom for instance: the famous real action RPG cave incident had been on the news. early 80's?). When, several years later, I began to find groups of players, I noticed something especially odd about them, as in poor wardrobe coordination choices, novel approaches to personal hygiene, etc. You know what I mean, or you know who you are....(joke). In retrospect, I understand that while I'm a geek at heart, I'm more of an artsy type in sensibilities. Long story short, I've never quite felt comfortable with the dedicated gamers who played face to face pencil and paper RPG. By the time playable (as in post Ultima Online) MMORPG hit the scene, I had grown a bit long in the ears and now I was no longer capable of enjoying group play with a group of frustrated teens (nothing personal, I was there once too). So, even as WOWC steadily encroaches the mainstream market, I'm still a single player RPG player. Naturally, being as obstinate as you can see I am from the above, I'm never quite satisfied with the RPGs I play, and usually I can't finish the main quest or story line of any game I buy without turning to modding it as my primary source of entertainment. So, recently with the release of the G.E.C.K. for the Gamebryo based Fallout 3 from Bethesda, arguably the best modding tool ever created to date, I have been hard at it for literally weeks of man hours and finally beginning to think, hey, if I have this much dedication and concentration, why don't I just roll my own? Thus began a journey into the world of game engines which finds me here today making my first blog post as a brand spanking new Torque Owner.

So, I want to make an RPG. My goal is to create something that <b>I</b> would enjoy playing till the finish, or lack of finish if I make it open ended, or many finishes if I try something like Neverwinter Nights. I know, I know, the scope of the task is of an order of magnitude beyond an FPS, which, despite the great tools and engines available, is still months of work.

I know I am not alone, from the very fact that there is a stereotype of the starry eyed noob who wants to create the world's biggest RPG etc. etc. Therefore, I've decided to blog my experiences as I go along. No guarantee that I won't drop this and move on to something else tomorrow, next week, or next year, but for the meantime, if you are in any way like minded, follow along with the story.

Before I get to the point of the post, specifically the title above, a note about aspirations. What don't I like about single player RPGs that leads me to want to build my own? Well, there are the obvious little limitations that always result from any project which is limited in scope by budget restraints and sales goals. When making money is priority number one, compromises must always be made, and thus we have a world where a game like Fallout 3, which from a idealist player's perspective (me) has plenty of glaring flaws, can be acclaimed to the extent that it has. Because the industry and the players know this was done on a budget, on time. Amazing. As an aside, I know their secret which is that, important engine changes and great visuals aside, it is basically Oblivion re-staged in post apocalypse Washington DC which complete content replacement but not as many low level changes as you might expect. I know this from modding both games and noticing that many Oblivion specific script functions remain, albeit untested, in Fallout 3 (horse riding related stuff for instance). Anyway, the limitations are too numerous to mention.

The biggest problem I have with current RPGs is the very fact that they tend to be based on linear story lines, which obviates <b>scripted AI</b>. Even in Fallout 3, with its amazing scope and non-linearity, the minute you play outside the scope of their pre scripted quest concept, you begin to see through the cracks to the other side of the facade and it all starts to look like Disney Land. Ai behavior becomes weird, bugs appear, etc.

As a single player RPG fan, this is fatal. I need the AI to behave naturally, as a typical human might, almost no matter what I do. This obviates the discarding of straightforward scripted AI to be replaced with all of the bells and whistles of embodied, reactive AI (I'm reading the book, more on what that means in the future if you don't understand it). With torque, this probably means digging into the source code, but as a learning experience, I'm game.

So there, I have further upped the anti. Not only do I not have a design, a specific goal, or a true understanding of the scope, but I even want to implement true AI. What a noob. (Edit: not true AI but "believable" AI, thanks to Brandon for pointing that out)

So what can I do? Iterative design. And we come to the point.

We've all heard of iterative design, reading forums, books, or actually doing it. Some of us have heard of or studied or done Xtreme programming.

Basically, when I say iterative here, I borrow a bit from Xtreme programming and define it as building only what you need when you need it, and then repeating the process, building more on top of what you built before, scratching and rebuilding older bits as required. This means that you define the first step you make in very narrow, achievable terms, and get right to work on implementing only that. It also means, again from Xtreme programming, that you don't get too concerned about the end goal yet, as that is likely to change anyway. This comes from the Web development world where the customer is always changing their requirements three months in. In the past, the programmers would say *&%` you! and the lawyers would be brought in. With Xtreme programming, they say, OK, well, so far we've only built this much, so we only have to rewrite this much. The point of setting small goals, or steps, and I'm not explaining this well, is that the result of each step must be something that you can actually run and interact with. So, for our RPG, my first step will result in a tiny level, with a player character, for instance. No inventory, no items to pick up, no NPC's with associated AI, no triggers, no portals, no nothing. Just a character you can control to walk around an interior level. That is a very achievable first step, or iteration.

The next iteration would be to, say, add a portal and an extension to the map beyond that portal, in which there may be a trigger, or an item to pick up, or an NPC, each of which would require the associated underlying systems such as inventory or AI. But even then, the AI would only be required to perform the small tasks that that one specific NPC wants to do, or the inventory would only need to handle the one item. Provided you script this functionality in an extensible object oriented manner, than rewriting can be kept to a minimum and you can do more re-factoring and adding on of code rather than scrapping and rewriting as you add complexity. Again, more on that in the future.

Perhaps Xtreme programming is not an option when your goal is to develop a commercial game in time for the such and such trade show and make a profit, but it is perhaps the ideal way to go for a hobbyist (that's me) or small team which is learning as they go.

That's all for today!

Chris.