This month I was far more productive with my still-not-named episodic RPG project – Using two monitors makes life so much nicer.

This month also marks a special occasion as I’ve now been working on the project in my spare time for one whole year now.

Time sure flies..

This month I’ve added features for NPCs to the map/world editor and spent a little time working on the datastructure for NPC dialog trees. The thing with dialog trees is that they’re not actually trees – But rather directed graphs. When you think of modeling conversations as a series of nodes which represent the state the conversation is in and a set of response options which represent a transition function to another state in the conversation, it’s actually quite simple.

I’ll be using lua scripts for the dialogs, that way I can easily add conditional responses – For example you wouldn’t want to accuse someone of murder unless you had gathered enough evidence to support the claim.

Also I’ve decided on doing a ‘pilot episode’ which will probably be a 3-4 hour long game to test the engine and actually produce a finished product before the turn of the century. I’ve sketched out some details for a side-story which will take place in the same world as the other episodes, but will involve a different protagonist.

See you next month.

Its not quite the end of the month, but it’s close enough to write my monthly review on the rQuest project.

Work this month was slow as I’m starting to pick up other (non-computer or only semi-computer related) projects / hobbies to do in the evenings.

Programming a video game after spending 8 hours a day programming line of business applications can be a little hard on the brain, eyes and body. Getting exercise and spending time away from the computer to degauss in the evenings can be really satisfying. Realistically I can only commit approx 8 hours per week for rQuest, but thats ok – as it’s a non-zero amount and it’s surprising what you can acomplish in an on-the-ball-ready-and-focused 8 hours.

This month I created the CRUD to add items to the project – For all your wands and shields and swords of much smiting +12 against trolls.

Screenshot Below

This was actually one of the more boring tasks to do as it is stupidly similar to what most database front-end working programmers do day in and day out – Someone has a product and it has these data fields which mean stuff and an end-user needs to enter new products and edit existing products etc…

Basically if I was running a medieval shop (or is that shoppe?) this months work would have been scarily similar to the functional specification of that shop.

I also did a little bit of story development. So the ‘soft’ areas of the game, if still neglected haven’t exactly been forgotten about.

As I have worked on the game editor it has become pretty obvious that I need to get another monitor – Often I need to run two IDEs (Visual Studio for C#, CodeBlocks for C++) with perhaps a paint program, or an Internet browser open. (Three monitors would be cooler, but my budget doesn’t extend that far just yet.) It’s almost an exercise in inefficiency to keep developing on only a single monitor.

So next month it’s dual monitors!

March was a pretty slow month for me.

Most work this month was spent re-organising the level editor, which is slowly transitioning into a complete game editor.

Instead of creating maps, you now create a project – Which contains multiple maps and each project has its own data for monsters, npcs, scripts and tilesets. This is largely a bi-product of deciding to go episodic, but it also centralises much of the workflow involved in creating a game.

This process will take a while as it’s much more complicated to write an application that manages n things when you only originally planned to handle one thing.

So work this month will probably continue to improve on the game editor. But I should really get around to getting the C++ game engine UI stuff implemented sometime.

Also, does anyone know of a (Preferably Cheap) syntax highlighting edit box for .Net? – I want to add the ability to edit LUA scripts into the game editor and would like to have some syntax highlighting.

The other question I was pondering with respect to my Indie RPG was how to deal with maps.

I thought that implementing an Ultima7 style continuously scrolling seamless map would be cool as there are not many tile-based RPGs that do that sort of thing. I even sketched out an algorithm that was based on a tree that would grow and store neighbouring maps in memory in order to stitch them together at render time, and I’m pretty confidant it would work quite well.

However, I’m not going to do it.

And the reason for that is because I’ve been thinking about the story line and other aspects of my game and it’s now going to be delivered in a series of episodes!

Having a seamless map kind of makes sense if you are planning a huge world, but as each episode is going to be its own game having the traditional style go to the edge of the map and load a new area makes more sense.

Delivering a series of episodes means smaller development cycles and a tighter focus per game. I think I’d rather play shorter games as I mentioned in Too Many Games to Play

Now that I’m done thinking about things, I can go forth and actually start producing some real work!

So a little over a week ago I posted about two fairly important decisions I had to make about my Indie RPG and after pontificating over a couple of hundreds of dozens of cups of tea, I’ve reached my decisions.

Decision One : The Resolution

Will be 1024×768 or 1024×700 in windowed mode – The extra 68 pixels in fullscreen mode will contain a little border graphic, the reason for a difference between windowed and full screen is because if you are playing on a laptop like mine and you only have 800 vertical pixels to play with and approx 80 of those are taken up by the taskbar and the application window, in order to fit the game on the screen, you have to cull some pixels.

This means that the game will look large enough to be played in windowed mode, so people won’t feel the need to go to full screen to get a more immersive experience and my old school pixel art won’t look all blurry when the game is played at non-native resolutions. Of course, if you were really old school you’d have a CRT monitor anyhow.

I may also offer a widescreen mode down the line, but I’m not promising anything at the moment.

1024×768 also allows me to have larger, easier to read text on the screen which is good for older more experienced people

The tilesize is going to remain at 32×32 at least for now. I’m going to make use of some stock art assets found around the internet and in the RPG Maker community for now so I can begin to piece togeather the game without having to fork up heaps for custom art assests straight away.

No screenies this month, and possibly not for a wee while – I’ve mainly been writing code.

This month I worked more on the level editor – You can now place monsters in a map and edit the walkable areas (as well as set tiles to have a default walk/no-walk flag, which pretty much means you don’t have to set no-walk areas at all…)

I also did more work on the C++ game engine side of things. Most of the engine is a port of the Java class structure from Caverns of Underkeep – With some improvements and refactorings and taking into account that this game is going to be more data-driven.

So at the beginning of the month I did what I almost never do and wrote a class diagram on a scrap piece of paper, showing inheritance trees and associations between objects, I’m sure my diagram doesn’t follow the UML spec what-so-ever, but it was helpful enough for me to start fleshing out the header files, and providing some basic implementation for the classes that I was going to be working with.

The feature that took me the longest to implement was remembering how to load data in C++ and convert it into some meaningful structures that the game can use. This turned out to be a pain, Esp. as I use non-null terminated strings of arbitrary length in many places – And I’ve mainly been working with garbage collected languages for the last 3 or 4 years.

So I spent quite a while in the debugger that day.

For example:

• When reading a file do I use .read() or .get() on an input stream. Which method increments the read position pointer automatically and which one doesn’t? (Hint: The first one I used didn’t.)
• If I create an object on the stack and then add it to a collection, what happens when the object goes out of scope? (Does it make a difference if it is a value type? Should I declare it on the heap? Should I store a pointer or a copy?)
• Why aren’t variables initalised to reasonable values, such as NULL for objects and 0 for value types?
• How do I create a dynamic multi-dimensional array that I can resize? (Hint: ObjectName **myArray) – And how am I supposed to clean up the memory when I resize it?
• Why are strings such a pain? How can I easily convert from char * myCharString to String myString? And back again?

So it probably took me about an hour to write the save/load map functions in the C# map editor and about 8 hours to write the equivalent feature in the C++ engine. (Too be fair, if I used C++  instead of C# every day at my day job, I’m sure the converse would be the case.)

That said, the C++ engine can now display maps created in the C# map editor, which is pretty cool.

Next month will probably be working on UI components and maybe getting some basic AI/Combat in place.

So I’ve decided to do these monthly progress report things. Weekly/Fortnightly seems to be too often, but if I do an end-of-the-month report I should at least have something of interest to talk about.

This month I finally got the project in source control – if you don’t currently use an online versioning system for your own projects I highly recommend getting EVERYTHING into source control NOW. I use assembla, but there are lots of online version control hosts out there, some better/cheaper than others.

I got a little distracted this month with recovering from the holidays, attending the 2009 Global Game Jam, playing Persona 3 (55 hours in, haven’t finished yet) – And just general working out what the heck I’m trying to build here.

The goal is now to get a working prototype done by the end of July this year (thats six months!) – I don’t know if it will be a public prototype yet. I might just show it amongst friends and collegues, but the idea is that I have a playable, if not complete game – I might take a leaf from Jay Barnson‘s book and try to make it a ‘pilot episode’ – We’ll see.

Most of this month and the coming month will be spent on the level editor and CRUD  – That’s Create Read Update and Delete for all you non-geeks. Basically all the stuff that you want to do when you want to deal with data. I’m using custom flat file formats in order to maintain portablility. Its the type of work that is not very hard but it is very time consuming as I have to sketch out the data format, write save/load functions and drag and drop UI controls onto a form and bind it to the data so that the user can edit and tweak all the values.

And RPGs have lots of data, oodles of it in fact. All those HPs and MPs and Spells and Monsters and Resistances and To Hit scores, and classes and races and lions and tigers and bears, oh my! Can produce some pretty large tables.

Heres a screenshot of the current world editor in progress

Now that we are full tilt into 2009 I’ve managed to find some spare time to work on the still badly named rQuest project.

I’m developing the map editor, which will over time evolve into the game’s complete authoring kit – With tools for dialog trees, weapon/monster database editing, import/export tools, all that fancy stuff.

I’ve started writing the application in C# with .net 3.5 and the GDI – If anyone has experience with the GDI you will know that it’s dog slow. I coded my own blit routine using unsafe pointers, and its about 10x faster than the .drawImage() routine, but its still pretty slow. Thankfully I’m only making the editor not the game with the GDI.

But its funny just how long a process it takes to get the software to ‘feel’ right. Simply creating software that works is easy by comparison – when I was testing out the map editor, it quickly occured to me that I really needed an undo. Which is a pretty non-trivial feature to add (especially for unlimited undo) Other things that we’ve come to expect in modern software also need to be implemented such as copy and paste. (Both in and out of the application.)

Another change I’m going to add is that instead of thinking in layers, which in some other RPG map editing programs cause me to get confused as to which layer I’m currently trying to edit. Each tile is going to be a stack – Left click adds a tile to the stack and Right click removes the tile from the top of the stack. Where the stack is capped at some height n. This should give appropriate visual feedback to the user as to what part of the map they can edit at any given time.

So work on rQuest continues, but at a slow and steady pace.