Carl's Notes

Well, I'd said that I was going to post information about my event system for my Tetris-Power project, so here it is.

The main idea behind this is a simple means for defining and detecting achievements.
If everything done in the game can be represented by a large list of events, you can simply scan the list for a matching sequence to see if an achievement has been matched.

In this case, everything really does mean everything, including keyboard commands, block motions, rows filled, rows removed, scores added, the advancement of time, error and warning messages for the log, and so on and so forth.

The overall structure of the modules then is heavily involved with the event system, with messages getting passed to the events system instead of directly calling other system, and each system that needs to read messages registers with the event system for the types of messages they're interested in.
(for example, if the logging system were enabled, it would read log_info, log_warning, and log_error message types)

This should greatly simplify the overall structure of the code so that not everything would need to know about everything else.
One of the main issues I have right now is that too many systems need to know about each other, which overly complicates my code.

So far I haven't actually added any code yet, as I'm still working on the concept itself.
Once I have that figured out, I'll have to rewrite large swathes of code to use the events system instead of directly calling.
Supposedly, it should be possible to gradually move over to the events system, starting with adding in code to trigger events, having each module listen for appropriate events, actually using the event listeners instead of the function calls as used now, then removing the functions used previously, then refactoring code to simplify things.

I hope it doesn't break while I switch over, but you can't tell until you start what might happen.

So I tried working on my Tetris-Power project again, only to find out why I'd dropped it last time: It was full of bugs.
(if it was literally full of bugs, it'd be easier to fix)

Something I'd done when upgrading the font system seems to have caused major bugs to show up elsewhere in the code, and even rolling back the font system doesn't seem to have fixed everything.

I'm racking my brains trying to find out what the problem could possibly be, but I'm not having much success just yet.

If I can remember how, I'll try to roll it back several versions to see where I went wrong, but chances aren't so good that it'll be that easy to fix.
(plus, there are some beneficial changes I've made since then, so there's no telling if I'll be able to undo all the damage I've done to it)

Though if I can't get anywhere with this, I'd hesitate a lot before I rewrite the whole thing from scratch...
(maybe I'll remember to put in more debug stuff and lots more comments)

---

Incidentally, I'm starting to work on my graduation application so I can get out of university for a bit.
Currently, I'm working on an appeal to use an arts course for one of my liberal arts credits, and need to find information dating all the way back to the Spring of 2005, which was less than a year after I got my very first "proper" computer, which means that I didn't have a good system set up for making backups.
(I'd had an old 286 with 1MB of RAM running Windows 3.11 and something like MS-DOS 6, so an upgrade to Windows 98SE was a huge jump)

However, I've still found time to try fixing Kathy's site under Internet Explorer (after about six months to a year, that is), and I just might be onto something, I'm just not sure if it works completely.

So, as I'm bored today (at least until I start making dinner, then going over to Grandpa's to watch the hockey game), I've decided to rewrite my font engine.

The current font engine I have for Tetris-Power is strictly based on bitmaps which are used to directly affect the screen pixels, which means that if I need a different font size, I have to create, load and track another font file, which is not a recommended practice.

So, I need a polygon-based font engine that allows me to specify a size for each line of text without needing a new font for each point-size I need.

This will need to be almost a complete rewrite of the engine from the ground up, as my previous font engine used a vertically short and horizontally long image to represent the font, and graphics cards like square images to work with.

Another reason I need this rewrite is that when a row gets removed, the score you get floats up from the vanishing row, which means that I needed to figure out exactly where in pixels I had to start the score, when the rows exist in three dimensions. With the new engine, I'll be able to specify the location of the score text by simply taking the 3D position of the row and moving it closer to the viewer, so much simpler and with even less chance for mistakes.
(also, it makes graphical effects like spinning text SO much easier to do, and it even gives me anti-aliased text for nearly free, too!)

I thought I'd post some of what's on my current TODO list for the tetris project:

• Add a background:
  Should include a sky and some ground, but levels can be indicated by the following:
  • Trees
  • Bushes
  • Flowers
  • Hills
  • Clouds
  • Cliffs (towards later levels)
  • Waterfalls (usually off the sides of cliffs. May have some trouble animating them)
• Add an Options screen:
  I'm not yet sure what to put here, but something about background music, enabling sound effects, graphics detail, etc. would be a good idea.
• Add sound effects:
  A light thud for blocks landing gently, a slightly heavier thunk for blocks dropped from a higher height, some kind of explosion for when a row gets removed (or maybe a cool disappear sound, I'm not sure), etc.
• Add background music:
  A friend has offered to make me a Techno remix of the original Tetris background music, but you can be sure that if I include that, I'll provide an alternative option as well as the choice of turning it off...
• Add an in-game menu:
  To pause the game and let the user change settings without having to restart their game.
• Add levels of difficulty:
  Because one level isn't challenging enough.
• Add a highscore list:
  To record how well you've done. I'll probably use SQLite to keep things relatively simple.
• Tweak or rewrite the font engine:
  To add multiple brightness levels. Right now, fonts don't look that great and could use some extra zing.
• Write a shaders module:
  To allow me to use shaders on a variety of platforms and even on some lower-end systems, including low-end Intel graphics cards.

Current status is that things look and play fairly well, but certain aspects could be greatly improved, such as block textures (I'd like to use a bump map, but that requires I use a shader so far as I know, thus the need for a shaders module).
I also seem to have lost some of my enthusiasm for work on this project, partly due to finally fixing textures (now I don't have a big challenge staring me in the face, just a bunch of smaller ones).

I want something big and difficult! Not small and difficult!
(sounds weird, I know)

From the "Why aren't they doing that?" file:

Personal Radio Recorder:
Most of us are familiar with Personal Video Recorders (PVRs) that let you pause, rewind and record live TV, but doing the same thing for radio would be a Very Good Idea™! It would also be fairly easy, and wouldn't require much of an interface.
(you might get away with a programmable MP3 player with a radio built-in. And no, I don't think an iPod would suffice...)

---

In other news, I haven't done much in the progress of Tetris-Power, I've still been thinking about it.

In regards to indicating levels, I've been thinking about a cartoony landscape that rotates sideways to hide the old level scene and show a new one.
(imagine that the background is a very large disk who's center point is a bit below the screen)

I'm not sure I want to change how the blocks look yet, but I should probably do something about realism with some kind of shadows...
The simple kind of shadows are when you draw everything twice, the first time you draw surfaces, while the second time you change how drawn objects are transformed so they fall flat onto a plane, which requires that you know where your lights and your plane are and where they're pointed. The disadvantage of this technique is that it works fairly well for projecting shadows onto a flat surface only, and to shadow other surfaces requires rendering everything once for every plane or polygon you want shadows on!
The other kind of shadows also require drawing things twice, but the first time is for shadows (looking from the position of the light, so it works best for sun and spot lights) and the data from that is used to render the actual surfaces. Which would actually require me to figure out how to do that...

I'm currently figuring out exactly how to draw nicely curved objects with OpenGL, and so far I think that a semi-circle texture is the simplest way to go. So far with a test program, this has proven very satisfactory, though I'll need to fine-tune a few things on it...

This, however, brings me right back to the problem of textures not actually showing up in the project... D:
I may have to cut the project apart and sew it back together again, adding pieces until textures no longer appear, assuming they show up again in the first place...

Not fun.

I've cut out the loading system to see if it's causing some problems (so far nothing), so I'll look for other stuff that doesn't make sense (which could be a lot of stuff).

Anybody know OpenGL or know somebody who knows OpenGL that would like to contribute to an open-source project and likes debugging? Anybody? Anyone at all?

Yay! Rows can be removed now!
(though textures still don't work)

Now, on to animating the removal...

First, I now own a half-interest in a second-hand EeePC.
Tiny little thing (7-inch screen, 4GB drive), most of what I've been doing are updates (I finally found Synaptic, a Linux package manager), and trying to do web-browsing on the small screen.

If there'd been one available, I'd have chosen a 9-inch model, but anything is better than nothing.
(and a small, lightweight model like the EeePC is better than a four-pound, widescreen model for portability)

---

Anyway, so far I've tweaked how the grid stores blocks to make it easier to check when a row is full, swapped out the 118KB cube (1248 faces, 1250 vertices) for a 26-face, 24-vertex, 3.3KB version with texture coordinates in the hope that I can map a texture on it to simulate a more complex cube.

So far, however, all my texture efforts have been in vain.

The image is loading fine, but none of my efforts have been able to pass the texture data over to OpenGL, and most of the documentation I've found is slightly incomplete. The missing part is how to separate loading the image into OpenGL and rendering with it.

The confusing part is that the same function call is used to set up the texture for loading and use it for rendering, so I'm not sure if I'm calling it the right way each time.

Anybody have any suggestions for OpenGL texture tutorial or reference sites? I'm a little stumped...

One of the things I said I was going to work on next was position interpolation stuff.
(i.e., the stuff that makes things move smoothly)

I'm currently taking a course on Numerical Analysis (basically doing math on computers and how to make it easier and/or more accurate).

What are the odds that I just learned about doing this sort of thing this last week?

Very high, apparently.

I'm using a variation of what's called Bezier curves, equations that are governed by two endpoints and two control points that determine the shape of the curve, and generate a smooth curve between the endpoints.
It's a variation because the original definition works in two dimensions, but the two sets of equations do not depend on each other, so are completely independent and can be used separately.
This simplified the code somewhat, as I only had to deal with one curve at a time, so could just use two instances to control the two dimensions separately.

Testing has been very satisfactory, as any change in direction is taken into account when generating the new equation set, by taking the current position and velocity and using them as the starting point for the next equations.

Overall, I'm not quite sure what's next. Several things need doing, such as the grid system (potentially a lot of work) and designing and displaying the backdrop behind and around the game-grid (needs a lot of thought).

Sometime after that, I need to redesign how the buttons, progressbars, etc., look, start work on the power-up system, add scoring and timing (including deciding on game modes. Especially the part about endurance versus timed modes, and if I should just include them all), all the particle and otherwise effects, then the final graphics redesign.

I should probably also stop working until midnight because I was up late due to working late the night before...

I've got another image to present:

Cafe Wall:


There's a few things about this I want to do over (such as the doorway itself and the glass on the lamps), but most of the scene is pretty good.
(kind of surprising for a scene that could be as old as two years, I'm not sure)

Anyway, I think this scene was supposed to be a backdrop for the cafe chair scene I did around that time, but now I'm not sure it would be a good fit...

Any comments or suggestions?

---

More news of Tetris-Power:

I've started on the underlying game graphics and mechanics, which translates to loading and rendering a Wavefront Object file.
Next is drawing four of them in a Tetris shape, then moving them with the keyboard, then dropping them into the grid, then, well, you get the idea.

Lots left to do, but it's definitely coming along nicely.

I have to say that I'm almost enjoying this (actually enjoying this would be getting paid to work on it), the amount of thought I have to put into it is a bit much sometimes, but then you realize that you're over-engineering it again, and you really shouldn't be doing it that way.
(kind of like when I made a tape into MP3s and the software crashed. I took almost four hours on Sunday almost tearing my hair out trying to recover the backup files when I could have easily re-recorded it again in 2 hours while doing other stuff as well!)

*insert momentous occasion music here*

Simple Tetris version 1.1 is now available for download!

The requisite screenshot:


It retains all features of the previous version, but now with greater stability.

Only one known bug has been introduced during this rewrite, and I'm not sure how to handle it.
Previously, you could hold down a key and the block would move repeatedly (for example, holding down the 1 or 3 keys would have the block spin continuously), whereas now you have the press the key repeatedly yourself.
(admittedly a minor bug, but a bug nonetheless)

If you do choose to download and play this, please let me know so I have some idea of how popular it is.
(and any feedback would be appreciated)

---

The improvements I have planned for Tetris Power are as follows:

• A "screen" system so I can show overlays, dialogs, etc. such as intro, option and help screens.
• Hopefully an improved animation system so I can animate almost everything.
• A more intuitive default key layout, with the option of changing them to fit the way you play.
• Colourful backgrounds to add that professional touch.
• Support for more block shapes, including three- and five-piece blocks, with options to enable and disable them.
• More bells, whistles and shinies.

We'll see how long it takes to complete...