My name is Vernon Dunmore. I am one of four members of the team Codus Vivendi.

I spent the last year of my life at Full Sail Real World Education, in the Game Design and Development Associate Degree program. The last three months of the program are devoted to a final game project. Students break into groups of 3 – 6 and create a game of their own design. Classes are scheduled from 9-5 three days a week, and the rest of the scheduling is left to the individual groups.

The game we created is a cross between an MMORPG and an FPS. We wanted to take the intensity of the FPS and give it the substance and ambience that is associated with the MMO. The classification we came up with to describe this cross-genre game is MORF³ , which stands for Multiplayer Online Role-playing First-person Fantasy Fighter. We named the game Aeons of Strife. The following is a journal of the events that occurred over the course of our development.

12:14 AM 09/30/2004

Today was the first Asset Production & Design class, and first presentation of our game projects. There were a lot of racers, some RTS games, a flight sim, a Smash Brothers Fighter, and us. Our enthusiasm for our game showed during our presentation of our idea -- we touted our wish list features (B List) as core features.

As a result, when we presented our idea there was a noticeable doubt by the teachers, TAs, and students of our expectations and capacity. And by 'doubt', I mean that Rob Garfield (the APD course director) loaded up an imaginary hand-gun and pantomimed blowing his brains out the moments after we said the word “MMO”. I did damage control by stating melee, magic and ranged combat were our core elements. I also said that we would add the fanciness once that was finished, but couldn't tell how well that went over.

After class I had a chat with Rob and I reiterated how we planned on prioritizing the tasks that lay ahead of us. He suggested that I make a list of the minimum amount of game elements needed in order of importance. We called this mini-maxi-hierarchization. Don't bother looking it up, it's not a real word.

It seemed to me that the members of Codus Vivendi all had different ideas of what our game will become. After class we clarified the game vision and redefined the core components necessary to finish. I plan to be able to refer to that throughout the course of the project.

5:23 PM 10/4/2004

I went over to Jeff's place to talk with him about the technical specifications. He is a very adept programmer, however at this stage he needs a little motivation to get started. We outlined in even greater detail the exact order in which tasks need to be accomplished in order to create our core game. Our next move consists of nailing down the APIs and the software we will use.  We also need to design more game elements (as we did today), and prioritize the details of our core technology.

5:00 PM 10/5/2004

We started to assign group roles today during an in-class meeting. The tasks are well fit for each member; we are all doing things relative to the capacities that we're interested in doing once we get out in field.

Group Roles

Project Lead: Vernon Dunmore
Tech Lead: Jeffrey Houde
Design Doc Manager: Vernon Dunmore
Tech Doc Manager: Jeffrey Houde
Asset Director: Vernon Dunmore

Czar Roles

Render Czar: Jeffrey Houde
Audio / Physics / Collision Czar: Petrus Bonang
Networking / Interface Czar: Jeff Marr
AI / Gameplay Czar: Vernon Dunmore

The group roles are pretty straightforward, so I won't explain them any further. The Czar system is a more specific group hierarchy that states that a specific person is in charge of a particular section of the game. Furthermore, the Czar makes the final decisions about the implementation of his own module, since he is going to be the primary person writing it.

We also began a walkthrough for the game; what would occur, when, how, etc.

12:56 PM 10/7/2004

Slowly realizing that our design documents aren't being produced as fast as they need to be. We are slightly behind on some, but since I've never made one before, I can't be sure. The course syllabus has deadlines for components of the doc, but the amount of detail I need to put into it for my particular project is still unclear. I'm just going to put everything I can think of into it to leave no room for any oversights.

I have taken the brunt of the DD because it interests me, and because I want the rest of the group to have the time to work on the parts of their modules that don't really require heavy design. This is what we did last time in SGP, and I feel I caught it earlier than before, but it still concerns me. I will give some tasks to some people that relate closely to the work they will eventually do. Beyond that though, I want to be certain they can program. I am just going to design the rest of the game DD on my own, as soon as possible. Switching between tech and design mindsets is draining.

5:14 PM 10/11/2004

We are now later into the 'month', with sufficient documentation, but not superfluous documentation. Tomorrow after class we will meet to cover the complete details of our game.

Spread myself too thin working on all document aspects. I gave code-relative DD tasks to others, and am waiting on their completion.

Additions to the statistics system were made. Writing the spell / skill learning is a big dependency, that definitely needs to work, and soon.

JeffH is having problems getting animation in RenderWare to work, our solution is asking Troy, (a TA here who wrote the base RW wrapper that we use), how to go about fixing it.

Also, JeffM cannot get DirectPlay to work. I am going to look for information / tutorials about the API to help keep him going.

PeterB has no work to do, since NovodeX requires animation capabilities from JeffH. The current solution is to give him the Component Interaction Matrix, since he is covering our physics and collision detection, it only makes sense that he draw up the matrix.

I have most of the design document to myself, but it is disorganized and discontinuous. The solution for me is to leave myself time to polish details and have a competent non-group member look it over.

3:27 PM 10/15/2004

At a meeting last night, we got a small amount done. Half of the time the group seems unmotivated; I realize I need to create a sense of urgency, but don't know how to go about doing so.

Peter is apparently an artist, and a good one at that. I just asked the group if anyone could draw concept art and he volunteered himself. Most of our class sees him as shy, but I think he's just not vocally loud. But anyways, 'shy' people are not shy when they are with an individual or group of individuals that make them comfortable. I think that as a whole, we make him feel comfortable. His art is going to be very useful in terms of asset production. He grabs the intentions of my 'programmer art' and makes it look nice enough so that the artists don't vomit while looking at it.

12:45 AM 10/17/2004

JeffH took the initiative to continue work on the flowcharts and modules in the tech doc without being told to do so. JeffM is looking into the way NetZ handles things, and plans to incorporate some of their methods into his DPlay code. Peter began work today on a model that will map death animations through NovodeX to RenderWare. He has this week to do it, after that I need him to begin Collision Detection wrappers and physics.

A number of scattered arguments have occurred over the past week. In the past they've gotten to me. I wonder how a group can be unified if so much time is spent bickering. Some days are good, some days are bad, nothing more. I can't let one day make the project for me, or anyone I am responsible for.

12:50 AM 10/18/2004

It is early, but we're trying to keep the memory requirements low for our game. Since we have a limited number of PC models, we're looking into putting models into bins for rendering. RenderWare has some nice features that let you copy and reuse a model residing in memory, instead of reloading it. The Memory Decrease from using one load per model, to one model being cloned is 195K -> 47K.

Also, I'm trying out a different meeting format. I go through each member and call attention from people only when their issues are discussed. This way keeps people informed, but doesn't bombard them with things that don't relate to their code. This keeps unnecessary 'events' from happening.

I started formally scheduling the tasks for our project today. I always dislike having someone estimate the time it takes to do something they've never done before. The best thing appears to be this:

1. Try to keep times low
2. Keep an eye out for overestimation
3. Allow those who overestimate by safety to do what they will - but make suggestions,
4. Granularize tasks as much as you can. Of the two:
   
   Write Paper ( 5 days )
   vs.
   Research Topic ( 1 day ), Write Outline ( .5 day ), Write Intro / Conclusion ( 1 day ), Write Body ( 2.5 days ) 
   
   which task is easier to determine when it's behind?

10:28 AM 10/26/2004

A major risk came in the form of the network. Not the API being used, but the client / server setup (that means JeffH and not JeffM). Initially, JeffH wanted to make a zone server and a gateway server, in an MMO setup, shown in the diagram below. Since our game is a cross between an MMO and an FPS, the rest of us thought that might be too slow to keep up with the fast – paced FPS action we want in the game.

Today we had Tech Doc Sitdowns, which is when a TA (or two), and John Fairchild (the APD tech specialist) go over the tech document and point out the parts where they see holes in the technical design of our game.

We were told that the above setup would be too difficult to synchronize, and more importantly too slow. JeffH had a hard time letting go of his design, but after some heavy coaxing, we convinced him to handle all aspects of all regions on the same server. In the long run, this way is what's best for all of us in terms of finishing what we want in this project.

12:10 AM 10/30/2004

Client and Server are now connected. They transmit data, and the client loads a world specified by the server. A problem arose when the client tried to load a level asynchronously, but was resolved by just making it load outside of a thread. A bit of delay is caused by JeffH learning JeffM's network API. He has questions that he wouldn't ask if he wrote it. Fortunately, since they're roommates, Jeff can ask Jeff questions in person, as opposed to online.

JeffM is finally free to start working on input after fixing all of the network issues that were revealed during integration.

Assets are not in yet, and Peter is scheduled to start on Collision Detection. He can use generic shapes to start testing, but it will still prove more difficult to do without the real artwork. Our collision detection is core (obviously), and I am concerned with its completion so much that I've put it at the front. I considered rescheduling to have Peter do audio first, since test assets can be downloaded easily. We're scheduled to have collision and physics done by the Alpha milestone, and rescheduling Audio first would push those two to the brink of the deadline. I suspect that the schedule will take bumps over time and don't want to take that risk.

10:51 PM 11/1/2004

I am garnering confidence with the choices we made as a group early on. I am also making sure that the rest of CV knows that fact as well. Other groups (at least 2, by my numbers) currently do not have a working build. They have nothing showing onscreen.

We knew in the very beginning that our project would be a lot of work, and concluded that cutting a corner with RenderWare would be a good idea. A client model is now moving and updating on both DirectPlay server and client, and visible in JeffH's latest build. Peter is testing collisions with the 10/31 build, and I am drawing flowcharts for Npc Motion.

With my Asset Acquisition, Scheduling, and Argument Resolving, I feel I have a lot to do before the AI. However, I can't allow myself to be the weak link in this chain. I am also personally ensuring that the artwork for this game is exceptional, and beyond what anyone has ever seen come out of FullSail artists.

8:01 PM 11/4/2004

Yesterday I met with Zaid about the Female Chainmail Model. He tried to show me the file, but it was corrupted. _And_ he said he made a JPG, but it was apparently missing from the drive he used to give me the files. For the second time, he brought James with him, whose artwork is not what we're looking for right now. Zaid seems unreliable at the moment, but a flash of inspiration came to me today. I saw a spark of competition in his face when I told him that Rory, our Modeler, made a very good model. If necessary, I will play off of his competitive nature to get better and more consistent work from him. There doesn't seem to be any other way.

5:35 PM 11/5/2004

Earlier this week we found that some groups were going to lose members due to unforeseen circumstances. As a class we decided that some people would need to be picked up and added into other groups. CV lost no members, but after discussing it, we said we would be willing to take someone on to help us complete our B-Grade Features.

Given the list of potentials, we found a handful of people to be competent and compatible. While discussing with my group, I brought up the cliché of thinking that new people would expedite work, but that training and rescheduling could wind up dragging us down. By the end of the week some individuals married into other groups, and the stragglers merged into a new group. Ours remained unchanged.

I feel it is better than worse this way -- we don't have to sell someone new into our idea, or de-n00b anyone to our practices. But most importantly, we don't have to deal with another voice during heated debates.

5:39 PM 11/6/2004

After another check-up email, Zaid submitted the model, but I believe it is in Maya 6 format. Comparatively speaking, Rory is far more consistent and reliable. But we need more models, and fast. I have no problems with spending an hour of my time talking to someone outside of class if it means we get more content into our game.

8:30 PM 11/9/2004

I felt the need to re-inspire my group; to both remind us of what we're up against, and of what we're setting out to do. A lot of our classmates continue to have doubts about what we can do, and I reminded us of that fact. Most of the group was inspired, but one of us was not. His behavior would have been counterproductive if the rest of us were the types to allow that to happen. It is a bit strange, but by making the expectations of our classmates our competition I feel a renaissance of drive in my group as well as myself. Go figure.

7:49 PM 11/10/2004

Today was a very good day. It appears that the morale of the artists is affected by the work of the programmers, and the morale of the programmers is affected by the artists. Joel, The Art Director, came by and told us that our 2D Menus (JeffM) was queued to an artist that would start on it soon.

PeterB has test models as of Monday that he can use to visually test the collision detection algorithms that he has. And JeffH and I have the actual game model to use in our tests of our respective parts of the engine. We also spoke with Chops Washington about our spell & weapon models.

Near the end of the day, Erick Fenstermaker showed us his first pass of the Endres Pathway level. It looks good, and he is very excited about working on it.

Rory, our primary 3D Modeler, also seems comfortable with me personally, which I also like. I want to make the artists feel like they're not working, but more like getting paid to do what they want to do themselves. When you keep people feeling that way, they yield better work.

12:12 AM 11/18/2004

Today was an asset day. I spent the most of it exporting files from Maya6->5->RW.

Erick came by and showed us the forest minus the leaves. The normals of the leaves point to the sky, which means that you cannot see them if you look straight up from below. We thought up solutions to the problem, and came up with two. First, since the level is all outdoors, most faces are not back face culled, turning off back face culling for this level (or parts of it) might not be as bad as it would be in others. The second is adding in alternate faces for view from below. Both have reasons for why they work in this situation. We're asking him to make a small test level for us to test out our theories and make a decision on that basis.

2:39 PM 11/19/2004

Today was another asset day. There were no exports today, but I did have to glance over a first pass of the spell models. Chops hooked it up, each one looks nice. The fireball has planar aspects to it that won't render when directly behind, so it will have to change.

There was also some discrepancy between what we met about and what was in the actual asset list I sent him. I also had to give him a scaled model to be able to figure out the relative size of the projectiles that he was working on.

Today I put Rory on to the game Guilty Gear XX Reloaded. I designed our enemy similar to 'Faust' from that game. I really like how his transition from squatting to standing is so intimidating and disabling. I wanted the player to get that feeling the first time he attacks one.

I can finally get started on the Fight / Flee State, but I still need to write up some structures, and finish the flow charts so that I can just dive in without thinking about the dependencies.

9:58 PM 11/20/2004

Peter found another weird flaw today. The first was that the OBB's were oriented badly; the second is that getting the triangles from a world sector sometimes returns absolutely nothing. He's not sure whether the problem stems from NovodeX or RenderWare, but I told him I thought it was cool that he could even find a bug in a system that polished.

I put together a minor PowerPoint presentation today, just the basics. The ppt is complete with concept slides, expert slides, and things of that nature. We have an Alpha presentation on Wednesday.  We're looking pretty good for Alpha, Peter is working on our death animations, and that's our big  B  B+ feature.

3:09 AM 11/22/2004

We are closing in on our Alpha Build. Peter finished the Physics ahead of schedule; he's always good with that. He is working on NovodeX RagDolling stilll. I'm concerned that I am having him work on that too much. We still have yet to move on to audio. However, once he finishes that, and JeffM finishes the menu components, all of our core tech will be done.

The Engine runs at 90fps, which is ok, but that is without integrated AI, or anyone else in the world. I make note of this, because in addition to Project Lead, I also bear the title of Risk Assessor. However, I find it difficult to do so when my knowledge about the engine is limited to what I ask and decipher by reading the code of others. I try to keep myself informed though.

The client has a good connection, but I can see the server updates occurring as ticks in the game. It's not fluid, but it still looks nice. The level is lush and very RPG-esque. Our second level looks to be a cave, which to reinforce the FPS aspects with multiple levels, jumpoffs, etc. The level artist has no way of testing out his levels to see how they look in our game because he cannot output his Maya levels to RW. I think if he could do that the process would be expedited a lot, and he wouldn't need to wait 2 days for us to give him more work. He's very excited about the project, and I like that. I keep finding myself trying to find more work for everyone to do, since they're all ahead of schedule. But that's a good thing. I have rough concept art (as best as I can draw) in my Sketch Book for PeterB to make fancy. I plan to pass it by the group, and hand it off the the Level Artist in the evening.

12:21 PM 11/23/2004

Our Alpha build is due today. Late last night we finished integrating the majority of all of our Alpha expectations into the game. Peter is working with NovodeX still, and I have decided I want him to continue doing so, even though he won't make it in time for Alpha. He's worked hard on it, and if we're all putting this project on our resumes, he deserves to have this in the game. If time requires it, I'll schedule him to start the audio just to get that done, and then go back to attempting Ragdoll Physics again.

Last night was the first time the game felt like a game. The rudimentary (run here, run there) Articifial Intelligence looked moderately intelligent, the collision detection algorithms detected collisions, and the network....networked. With all that going on, JeffH and I joined the same server and began slaughtering a few defenseless NPCs. It was the kind of moment where you smile for so long that your face gets tired. It was then that JeffH said what we were both thinking; "Now it's starting to look like a real game". The collisions only checked PC Sword to NPCs, so we couldn't actually fight each other. Of course, that isn't to say that kept us from trying...

1:53 AM 11/24/2004

Alpha presentations occur in the morning. Earlier today, we put our perfect, spotless release build on the test machine at school that our games have to demo on and... it crashed. We ran the game on all of our laptops and found no errors. If we could run the game on the machines we use to develop tomorrow we would have very little to worry about. However, that is not the case.

As the designated Risk Assessor, I concluded that we needed a backup plan, and the group agreed. The error has some relation to the test machine having more than one NIC card. Since we cannot test on it tonight to see if the latest build works on that machine, we're going to make a safe version of the game. JeffH (Tech Lead) knows the engine best, and is currently making a dummy client which doubles as a server. If all else fails, it will be able to exhibit collision detection, ground clamping, input, 2d components and all the other stuff that will show some (not all) of the progress we made. Hopefully, 'all else' will not fail. Right now I need sleep...and then when I wake, I imagine I'll need some caffeine.

10:15 PM 11/24/2004

Alpha presentations were today. The classroom was filled with classmates, teachers, TAs, and a few unfamiliar faces that snuck in while the lights were out. The actual presentation part went fine. We all knew our slides and made decent transitions amongst ourselves. During the game demo part of the presentation, we came incredibly close to being completely 0wnt by Murphy's Law.

I said I didn't want to be dependent on the network for our game to work, and as luck would have it, the network was down today. We knew the server wouldn't be able to run on the test machine, and prepared for it by running a server on one of our laptops. However, with no network capabilities, and no way to run the server on the test machine, we had to drop to our last resort.

The test machine couldn't connect to any outbound ports, so we basically had a network game with no network. Right now, the server crashes if it does not find an outbound internet connection. And as we expected, it crashed today, much to our dismay. If it weren't for that quick backup plan that we made last night … we would have had absolutely nothing to show to the class other than a bloody front end menu. JeffH ran the single-player mode after the server crashed, and I made a seamless transition during my narration of the gameplay. We left a handful of would-be detractors only enough time to snicker at our temporary misfortune, but soon after their jaws hit the floor. The level was lush, the animations were seamless, and the trails were hypnotic.

Nonetheless, two bad things happened and we were prepared for the both of them. We went through the game presentation as though nothing happened and made note of the game basics, the weapon trails, and the time of day / fog effects. The lesson of the day is that Paranoia is good … sometimes.

We got the collision map from Erick Fenstermaker today. It's not exactly what we wanted, but we have no way of letting him test his levels without our game to check his progress. A while back, we considered making a sandbox for him to use. However, since he does not have RenderWare installed on his home machine he cannot export RW files on his own, thus rendering the rendering idea useless. To resolve the situation, we're making a diagram of where we want all the boundaries to be in our game.

Now I'm off to do more asset related work. We need game music badly, and our 2D Artist needs feedback on the buttons he created for us. Also, I said I would send some videos of our game out to my brother Jay (to use in our trailer video), but he will need some theme music to serve as background music for the intro video for as well. The log ends for now, but in all, today was a good day.

Diary: To Beta