Minimal Bitcoin Miner in C#
Bitcoin was introduced in 2009 as a digital, decentralized currency. In the Bitcoin network there’s no central authority. Instead a public ledger of every transaction is validated and shared using peer-to-peer technology. Recently there was a surprising spike in the value of Bitcoins, making me curious to find out more of the system behind.
Bitcoin transactions are grouped and confirmed by encoding them in a block and adding it to a blockchain that is protected by strong cryptography. The cryptographic calculations aren’t provided by a central authority – instead anyone with capable hardware and mining software installed can provide computational labor to help solve the next block.
When a block is solved the successful miner (or mining pool) earns a reward in Bitcoins. (Currently 25 BTC – worth about 2500€) This is quite an incentive and so Bitcoin mining evolved into a very competitive market. Advanced mining software uses GPU’s to compute hashes a hundred times faster then possible on CPUs and there is even dedicated hardware with hashrates in the Giga-range and a better Hashrate/power-consumption ratio. Nonetheless incredible amounts of electricity (~1 GW/h) are wasted on mining Bitcoins which is basically spend calculating zillion’s of SHA256 hashes. Currently the combined network of miners calculates roughly 80.000.000.000.000 Hashes per second. Crazy stuff.
Miniminer
If you are trying to understand how Bitcoin mining software works but can’t find a reference implementation that is minimal and easy to understand (like me, 2 days ago) here’s my contribution:
Miniminer is a simple CPU based Bitcoin Miner in C#. Only about 300 lines of code but fully functional, open source and uploaded on Github.
It uses the basic GETWORK protocol to connect to pools and mines at ~400Kh/s a second. That’s not fast enough for serious mining but enough to find a valid share eventually, submit it to the pool server and be happy about the fact that you understand every single line of code that made it happen!
Grove Script
In the past months I’ve been reading up on procedural generation and emergence. My reasearch led me to develop a scripting language to be able to interactively ‘program’ L-system based fractal animations. It’s meant to be easy to pick up and modify, so it might be a nice sandbox for people that would like to learn to program or just want to engage in a session of creative coding. It’s quite fun once you’ve gotten the hang of it!
L-Systems
I think every programmer has dabbled with Fractals at one point. I remember generating Mandelbrot images on the very first computer my family owned.
Last fall Josh Parnell pitched his project Limit Theory on Kickstarter. His claims sounded way too ambitious for a one-man project, yet he raised more then 350% of the funding goal. So why did people trust him to pull it off? His pitch video was featuring quite some cutting-edge procedural generation technology.
This project rekindled my interest on that topic. I found a pdf copy of the book “Algorithmic Beauty of Plants” and was intrigued. The basic concept of L-System is simple enough. It’s grammar based rewriting system, a theoretical framework for modelling the development of simple multicellular organisms, introduced 1968 by the Hungarian theoretical biologist and botanist Aristid Lindenmayer.
But the above book introduced me to more advanced variants of L-Systems, adding context sensitivity, randomness and parameters. The examples were quite impressive and my urge to experiment with it irresistible.
Usually L-Systems define a bunch of production rules that operate on strings of symbols. When a rule is applied it replaces a sequence of symbols in the string (the predecessor) by a different, typically more complex sequence (the successor). The more iterations you perform the longer and more complex the string of symbols becomes. Symbols are statically defined to do a very specific thing like rotating or moving a cursor to generate a rendering.
Imagine the above string of symbols were generated using L-Systems. F is defined to move forward, + to rotate left and – to rotate right.
This method of programming line based drawings by controlling a cursor through commands that move and rotate it relative to it’s current state is called Turtle Graphics. The programming language Logo designed for educational use first introduced it and it’s quite intuitive to grasp.
Grove Script
My experiments turned into a little scripting language to interactively ‘program’ L-system based fractals, called Grove Script. At first it seems like it does nothing special as it is based on the concept of using turtle graphics in conjunction with L-systems, too.
But, instead of statically defining an alphabet of symbols that the productions operate upon and that have a specific meaning when executed, in Grove Script the symbols refer to executable code. The key idea is that instead of programming everything like in your typical programming language or forfeiting the flexibility of procedural languages completely like in your typical rewriting system Grove Script brings both worlds together. It allows you to use rewriting rules to create complex sequences of procedure calls by applying simple rules to equally simple axioms.
And unlike most other script- or programming languages where writing and running the program are separate steps Grove Script is aiming to allow users to develop fractal renderings and animations in an fun, intuitive process: There’s no compiling or restarting required when you change the code. You get instant feedback on your changes. I let Bret Victor explain why that’s a good thing!
You’re free to use your favorite Text Editor to write Grove Script. But when the script you’re editing is running in the interpreter every changes you make will be instantly incorporated.
Productions
The most basic productions just replace a token with a with some replacement whenever it is encountered. Rules need not be simple, though. Grove Script supports:
- Context Sensitivity – Rule can require multiple tokens to be encountered in order.
- Parametric Tokens – Tokens can have a parameters attached.
- Conditions – Production trigger only if the supplied boolean expression evaluates to ‘true’
- Consequences – Rule-specific code can be executed whenever a production triggers. Great to raise some counter variables.
Instruction Set
As mentioned before I use the concept of Turtle Graphics in Grove Script. All pixels are drawn by the turtle, controlled by only a handful of straightforward commands. Besides commands to modify the turtles spatial state you can set the size, transparency and color of the line. That’s it.
There are other commands to save and restore the interpreter state on a stack – very important to generate branching structures. There’s an instruction to call subroutines or a list of subroutines. And of course those subroutine-lists are ideal to be generated with L-System-inspired productions. So you got a command to initialize a list with the axiom and one to apply a ruleset. Last but not least you can define variables through commands that can be used in algebraic and boolean expressions.
But all in all there are only a dozen commands you have to familiarize yourself with to unlock the full potential of Grove Script.
Macros
I found a way to handle all the usual control flow (conditional branching, loops) with 3 basic instructions that operate on code blocks marked using Tab indentation.
- Repeat - Continue with the first of all previous instructions of the same block depth
- Break – Skip all following instructions of the same or greater block depth
- Gate - When condition evaluates to false, skip all following commands of the same or greater depth, except other Gate instructions op’s of the same depth.
Using these control-flow instructions to the desired effect is a little counter-intuitive though. So to improve the usability I added macros that will unfold to the required operations when the interpreter parses the script. Macros blend well with instructions and improve readability. Being able to write code that uses familiar words like “while, until, repeat, if, else” hopefully appeals to people with and without programming background alike.
Thanks to macros you can write this…
repeat 4
out 'fooset _c, 4
gate _c > 0
out 'foo
set _c, _c-1
repeatAnimations
The interpreter is written in C++ and quite fast. It renders even complex fractals that require thousands of instruction calls in real-time. If it’s possible to render 60 frames per second it would be a waste to never change the output. By using the time() function in expressions you can generate fractals that change over time and render animations. Another important function in that regard is random() which provides you with pseudo random numbers. You can use the time() to calculate the random seed to be used by the PRNG and suddenly the randomness is under your control. It took me a while to figure out how to use these elements in conjunction with productions but once you’ve gotten the hang of it it’s quite easy to script quite impressive looking vector graphics that let you forget that everything you see is rendered with the plain olde Turtle Graphics approach.
Demonstration
Excuse the poor quality. I just captured the interpreter with Hypercam instead of rendering to video directly.
To experience Grove Script the way it’s meant to be you should download this package. It includes a precompiled interpreter, full source-code, documentation, examples and a half finished tutorial! Feel encouraged to try it for yourself: Start Grove.exe. Drag&Drop some scripts from the example folder to the application. You can edit the script while Grove executes it and it will adapt at run time. If you’ve got questions and can’t find answers in the Docs or from the tutorial just comment below!
There’s no reason this should not run on Mac or Linux. If you want to compile it you need the Cinder Framework - the rest is just plain standard C++.
And if you just want to know how Grove Script looks like without having to download and run executables from the internet here’s the listing that generates the animation from the above video:
//set some basic constant parameters to be used in expressions
set speed, 15
set fadetime, 20
set budstep, 5
set maxage, 70
set leafAge, 5
//let the size of the structure adapt with window size
set step, 0.5*_height/480
//these variables change every frame because they all depend on the current
//value of the time() function.
//Thanks to the modulo operate age wraps around at maxage.
set age, (speed*time()) % maxage
set fade, min(1, (maxage-age) / fadetime)
out fade
set age, 2.5*age^0.55
set frac, max(0.01, frac(age))
//each structure should look unique.
//So we use a different random seed for every new structure
out 'rnd seed, floor(time()*speed/(maxage))
shuffle floor(time()*speed/(maxage))
//init the turtle
dir random(360)
pos 0, step*(5*age-100)
//set the axiom of the 'plant' structre
seed plant, A(0,0,1)
//apply the ruleset 'r1' as often as the plant is old.
repeat age
grow plant, r1
//fun info: how many buds were generated?
out budCnt
//render 'plant' twice - mirrored.
push
rotate -45
run plant
pop
rotate 135
run plant
//Now the Tokens that the L-System operates upon are defined.
//light blue tips of the structure
#A(age, angle)
run shade(leafAge/frac)
size frac^0.5
move step*frac^0.5, angle
//what becomes of A in the next cycle - a simple curve
#B(t, angle)
run shade(leafAge/(t+frac))
size (t+frac)^0.5
move step*(t+frac)^0.5, angle
//B becomes C when it's old enough to spawn leaves - does the same thing, though
#C(t, angle)
run B(t, angle)
//a growing leaf
#L(t, angle)
run leaf(t+frac, angle)
//a finalized leaf
#xL(t, angle)
run leaf(t, angle)
//a growing bud
#Y(t, spread, angle)
run bud(0.3*(t+frac), (spread+frac)*budstep, 0.1*angle)
//a finalized bud
#xY(t, spread, angle)
run bud(0.3*(t+frac), (spread+frac)*budstep, 0.1*angle)
//render a bud
#bud(len, angle, curve)
run budshade(1/len)
//left side
push
size 2*len^0.3
rotate angle
move len^0.3, angle
size len/2
rotate -90
move len, -angle
pop
//right side
push
size 2*len^0.3
rotate -angle
move len^0.3, -angle
size len
rotate 90
move len, angle
pop
//stem
rotate curve
move 0.55*len^0.7
//render a leaf
#leaf(len, ang)
size 1
rotate 10*ang*len^0.6
move 0.5*len^0.3
move len^0.6, ang*(60+len*5)
rotate 180 - ang*0.4*(60+len*10)
move 0.5*len^0.5, -ang*(len*5)
move 0.7*len^0.5, ang*(len*10)
//color buds
#budshade(alpha)
visible alpha * fade
rgb 1.0-0.8*alpha*alpha, 0.6*alpha+0.2, 0.2*alpha*alpha
//color rest
#shade(alpha)
visible alpha * fade
rgb 1.0-0.5*alpha, 0.5*alpha+0.5, 0.3*alpha
#[
push
#]
pop
//The rules are evaluated in the order of appearance.
//rare rules appear first, if all are skipped the general case is usually to just grow the existing tokens
#r1
//BRANCHING
//higher c makes a vine less likely to bend or split but increases chance to spawn buds
A(i, j, c) : rnd(0, c) > 2+budCnt -> B(1, j) Y(0, 1, j)
raise budCnt, 1
//if old enough there's a chance to split into 3
A(i, j, c) : i > c and rnd() < 0.5/c -> B(1, j) [ A(0, rnd(-40,-60), c+2) ] [ A(0, rnd(40,60), c+2) ] A(0, 0, c+1)
//...or into 2
A(i, j, c) : i > c and rnd() < 0.5/c -> B(1, j) [ A(0, rnd(-40,-60), c+1) ] A(0, rnd(40,60), c+1)
//or just change the direction
A(i, j, c) -> B(1, j) A(i+1, j+rnd(-90, 90)/(c+1), c)
//LEAVES
//based on the curvature leaves spawn on different sides
B(i, j) : j < 0 and i = leafAge-1 -> C(i+1, j) [ L(0, 1) ]
B(i, j) : j > 0 and i = leafAge-1 -> C(i+1, j) [ L(0, -1) ]
//GROWTH
//and of course everything needs to grows
Y(i, j, k) -> Y(i+1, j+1, k) xY(i, j, k)
B(i, j) -> B(i+1, j)
C(i, j) -> C(i+1, j)
L(i, j) : i > 1+rnd(7) -> xL(i+1, j)
L(i, j) -> L(i+1, j)The following wall of text contains musings on how todays video games are missing their potential. I analyzes the game industry’s focus on surface values and why that’s a bad thing. I argue that photo-realism isn’t required for immersion, and point out what is instead. I draw parallels to the state of AI research. I critizise a lack of investments in R&D and a stagnation despite video games being far from realising their potential. Read it!
I wanted to create games for the better part of my life. I enjoyed playing video games but what fascinated me was their yet unfulfilled potential. Hardware was growing more powerful at a staggering pace and games where the only kind of consumer oriented software that would motivate people to upgrade their systems every two years. It was this synergy between hardware and video games that allowed both industries to prosper. We were witnessing the birth of a new medium and I wanted to have a part in shaping it.
Moore’s law, the prediction made in 1965 that the complexity of integrated circuits would double every two years, remained surprisingly accurate for over half a century. The improved capabilities of electronic devices changed our lives profoundly. When surfing the net with my smartphone, reading a book on my kindle, depending on amazon, google, wikipedia or the navigation assistant to solve most of my day-to-day problems I realize how technology has advanced way beyond what I could have imagined 15 years ago.
Video Games on the other hand turned into a disappointment. Not because my career plans failed – I have studied multimedia engineering and make my living in the games industry. But the product is not what 15-year-old me would have expected videogames to be like in 2013. Not remotely close. That’s a thought provoking revelation for someone that devotes most part of his waking hours to games.
Partial Occlusion Field-of-View
It seems like there are problems that I have to solve over and over again. Pathfinding is one of them. Another is visibility determination.
When you need to know if there’s an unobstructed line of sight between point A and point B you typically just cast a ray from A to B and see if it hits something. The set of all points that can be reached with an unobstructed line from A are referred to as A’s Field of View. Calculating it accurately would require an unlimited amount of rays, so you typically sacrifice accuracy by partitioning space in some discrete data structure. Grids, Trees, Pixels, Voxels… that kind of thing.
In this post I’ll focus on FoV calculations on square grids. Using a low resolution square grid is a pretty huge abstraction, and the whole point of this article is to show how to get some accuracy back without losing too much performance! 
But let’s start simple…
Calculate Field of View with Raycasting
As long as you only care for a Boolean result (is a square B visible from square A in the center) it’s straight forward to calculate using raycasting.
First of all consider how many rays you need: Just enough so that every field that could potentially be part of the Field of View is passed by at least one ray. So better define a max range or you would still have to cast an unlimited number of rays despite working on a low resolution square grid. Now select fields at maximum range so that those fields completely enclose your center. Draw lines to them and all closer fields are automatically tested too!
Raycasting only towards fields at max-range reduces the algorithms complexity from O(n³) to O(n²). Thtat’s a huge difference! Imagine how the illustration would look like if all cells would be target of a raycast…
For line-drawing you can just use Bresenham’s Line Algorithm. All clear fields covered by the line are part of the field of view. As soon as you hit an opaque field abort the raycast. It’s not part of the field of view and all other fields that follow along the direction of the ray are occluded by it.
There’s room for improvement
When I tried to use the raycasting approach to implement dynamic lighting for an isometric 2D game engine I found that I would have to find an improved algorithm to make it look great.
Consider an opaque field casting a shadow. A lot of fields will neither be fully lit nor fully in shadow. From light source’s point of view those fields are partially occluded! So I wanted my fields to specify their membership in the Field of View as a percentage (0 – 100%) and not just a Boolean (true or false).
The other limitation I hoped to overcome were integer coordinates. It didn’t suffice to say a light-source is on Field (x,y) – I wanted to specify the exact position on the field the light would come from. So I needed to be able to precisely define the center of my Field of View by using non-integer coordinates.
The Mission
I hope I managed to convey an idea of what problem the algorithm I’m about to present is trying to solve. To be sure let’s sum up the situation:
- We operate on a square grid that partitions a 2D plane in equally sized fields.
- Some squares are clear (will allow light to pass through) while others are opaque (will block light).
- We define a point for which we want to calculate the Field of View.
- We want to know the degree of occlusion for each square closer then the max range.
My research didn’t yield any established solution so I accepted the challenge to find a new one (or reinvent the wheel, who knows). Here’s a demonstration of my solution in action:
The full source of the demo is available under the MIT license. However, deducting an algorithm from poorly documented AS3 source code might be a little inconvenient if you want to understand the principles behind it. So I conclude the post with a explanation of the algorithm.
Partial Occlusion Field of View
It helps to think of center of the Field of View as a light source from which light is spreading equally in all directions. The spreading light will exit a square on the edges facing away from the light source and enter a neighbouring square. So for each square you can clearly identify one or two neighbouring cells that all the light is coming from.
Now all you have to do to know how much light is going through a particular square is to ask the relevant neighbours to describe their contribution. If a neighbouring square is opaque it will contribute no light. Otherwise it will contribute a portion of the light it itself received – which can be anything between zero and the width of the shared edge.
Circular sectors to represent beams of light
A simple way to express the “portion of light” is a circular sector. With the center known we just have to store two angles to define such a sector.
This illustration shows how the light passing between two occluders can be described by specifying two angles: Alpha and Theta.
Due to the nature of the grid one sector (e.g. two angles) per neighbour is enough to describe it’s contribution! This is because occluders are always one square in size and thus the shadows they cast can’t be smaller then one square. With this method of describing lightbeams it’s easy to form unions and intersections of lightbeams, too. Unions, when we want to combine the light coming from two different neighbouring squares and Intersections when we constrain the result to the size of our current square.
The Algorithm
The algorithm can be summed up as follows:
1: Identify all neighbours that are not opaque
AND share an edge AND are closer to the center.
2: Store the contribution from the first neighbour
as circular sector A.
3: IF a second neighbour exists: store the contribution
from the second neighbour as circular sector B.
4: Calculate a circular sector C that describes the maximum
amount of light that could possibly pass the current square.
5: Form the UNION of A and B and INTERSECT with C to receive D.
D describes the amount of light actually passing the current square.
The smaller the ratio of D to C the bigger the occlusion.
The following illustration demonstrates the process on a square that happens to be half occluded. Only one neighbour contributes light, while the other is opaque.
This animated illustration shows how to calculate the occlusion of a square based on the occlusion of the two of it’s neighbours closer to the center.
Calculating a Squares circular Sector
But how to calculate a circular sector that describes the maximum width light beam that could go through a particular square (labeled ‘C’ in the above illustration)?
To find it you simply calculate alpha for all four corners of the square and pick the largest and the smallest angle to form the circular sector. The resulting sector will enclose the full square.
The illustration shows that there’s a distinct pattern in what corners you would have to pick. So you don’t have to calculate four values and then discard two. Instead you can evaluate the sign of the x and y component of the vector pointing from lightsource to the square and to predict the correct ones. And if you’re going through the sourcecode of my reference implementation and notice these weird looking lines now you know what they do! dx and dy describe the position of the square in relation to the center and (x1, y1) and (x2, y2) are the corner’s that define the sector.
var sx:Number = (dx == 0) ? 0 : (dx < 0) ? -1 : 1;
var sy:Number = (dy == 0) ? 0 : (dy < 0) ? -1 : 1;
var ox:Number = (sy == 0) ? sx : 0
var oy:Number = (sx == 0) ? sy : 0
var x1:Number = x + 0.5 * (1 - sy + ox);
var y1:Number = y + 0.5 * (1 + sx + oy);
var x2:Number = x + 0.5 * (1 + sy + ox);
var y2:Number = y + 0.5 * (1 - sx + oy);Recursion vs Iteration
Another difficulty worth mentioning is that to calculate a particular square’s occlusion the occlusion of it’s relevant neighbours must be known first. How can we make sure they are already calculated?
The obvious approach would be to use recursion, which means that whenever we query the sector of a cell that hasn’t yet been evaluated it’s sector will be calculated before the query returns the correct value. To calculate the FoV (e.g. the degree of occlusion of all squares within max-range of the center) using recursion, trigger the evaluation in the four corners of the bounding rectangle. All squares enclosed by the rectangle will be visited before the calls return. There’s just no other way for the recursion to terminate than by reaching the center tile for which the occlusion is always known. This is when the evaluation-method ceases to calling itself on neighboring cells and begins to return values. It’s a good idea to implement it so that each cell is only calculated the first time it is queried and cache the result. That way subsequent queries don’t incur an unnecessary reevaluation.
There’s nothing wrong with that approach as long the stack space suffices. But if this could be a problem it’s also possible to devise an iterator pattern that will visit squares in such an order that all relevant neighbours have already been calculated.
The demonstration I’ve linked implements both the recursive and the iterator based version of the algorithm as well as some raycasting based FoV as a reference. Here’s the source code of the demo! It’s released under the MIT License and includes all relevant dependencies. I hope you enjoyed the lengthy read, feel free to post questions or comments!
Effortless Animations in Flash
You want to display an Animation in Flash. The animation consists of a number of frames, each is saved as a small image somewhere on your HD. You wonder: What is the best way to do it?
If this scenario applies to you, here is one possible answer that does not require the Flash IDE or any full blown framework. It involves a free tool and a couple of small classes (also free) that should be easy to integrate with your project.
Step 1: Create an Image Atlas out of the individual images.
What is an Image Atlas? A big image containing many small ones side by side. Also known as Texture Atlas or Sprite Sheet.
Why do we want one? Because it’s easier to embedd one image file than many. It’s also reducing the size of the SWF because it compresses better then individual images. We’ll use a tool called Texture Atlas Generator to create our image atlas.
- Check “PNG” so an image atlas is generated.
- Check “Pack Transparent”. This will pack the images as small as possible by removing empty pixels. The 17 original images have a resolution of 128 x 64 each. The image atlas has a resolution of 400 x 182. So the same animation takes up roughly half the amount of pixels.
- Check “TXA” to generate another file that stores meta information e.g. what frame went where in the image atlas.
- All other options can be un-checked.
If everything works you end up with 2 new files: A TXA file and a PNG image atlas that should look roughly like this:
Step 2: Embedd the image atlas and the meta data in your Flash.
As described in my other blog post you can embed all kind of data in a SWF.
[Embed(source = "../Resources/MoikSlashAni.txa", mimeType="application/octet-stream")]
private static var animationTxaResource:Class;
[Embed(source="../Resources/MoikSlashAni.png")]
public static var animationPngResource:Class;Step 3: Use the ImageAtlas and AnimatedBitmap class to display the embedded animation.
In the Code Sample you’ll find a couple of classes from my library of custom code. Feel free to use them for your own projects. In this tutorial we’ll use the ImageAtlas class to unpack the individual frames from the image atlas and the AnimatedBitmap class (that behaves just like a normal Bitmap class with some extra functionality) that we will add to the stage to display the animation.
//unpack the embedded resources
var imageData:BitmapData = (new animPngResource as Bitmap).bitmapData;
var metaData:ByteArray = new animTxaResource;
//create an image Atlas storing our animation
var imageAtlas:ImageAtlas = new ImageAtlas(imageData, metaData);
//create an animated bitmap...
var animation:AnimatedBitmap = new AnimatedBitmap(imageAtlas);
//...make it play a sequence of frames...
animation.assignFramesByName("MoikSlashAni_0000", 16);
animation.play();
//and add it to the stage
stage.addChild(animation);It’s as easy as that!
Download the sample project here: ImageAtlasTutorial
Step 4: Make it your own.
You probably want to have a look in the source code (to learn what AnimatedBitmap can do) and wrap the functionality with your own code. The AnimatedBitmap has a couple of functions that make it flexible to use.
- You can toggle wheter you want the animation to loop.
- You can set the current frame yourself or let it update automatically (by calling the play method).
- You can reassign new frames so you could store a lot of different animations in the same image atlas. Running, Slash, Jump and so on. So one image atlas and one animated bitmap could be enough to display a full character. You just have to change the sequence of frames you want currently played e.g.
assignFramesByName("jump001", 14)to play 14 frames of jumping. - You can pass a callback function as a parameter to the play that is called when the last frame is reached (and the animation isn’t looping) to play a follow up animtion.
I hope this tutorial has given you something to get started with! Feel free to post further questions in the comments.
Copyright Notice: The sprite I used in the demonstration is public domain and consider the code used in the sample as released under the MIT license.
Finally blogging on my own webspace!
I have a domain, webspace and unlimitted traffic… there is no good reason beyond lazyness to not host my own blog.
So, no more forwarding to wordpress.com for I found the time and bravery to make the move! I hope everything still works. In the days to come I’ll explore the new customization options that I now have.
Thanks to Barry from GamesChart.com for the nudge!
The Challenge
A contrary approach is to put all resource files on server and load them at run time. But different types of resources require different code to load, if you require many resources at once you have to open one connection for each resource (adding a lot of overhead) and you have to keep track of multiple loaders. Lastly many resource files are compressed efficiently when embedded into a SWF and it’s sad to miss out on that kind of optimization.
Our Solution
The Details
Matrix Magic
1. How do the properties a, b, c, d, tx and ty map to the coefficients of the matrix?
2. How to use the matrix to transform a Point?
- deltaTransformPoint transforms a point without applying the translation.
x' = a * x + c * y; y' = b * x + d * y;
- transformPoint transforms a point including the translation.
x' = a * x + c * y + tx; y' = b * x + d * y + ty;
3. How to setup the matrix to get the desired transformation?
If you need it you can set tx and ty to move the transformed points relative to the new origin.
And why do I blog about math, now?
Simulating Fluids in Flash
A thread in the FGL forums got me interested in the simulation of fluids. Of course this topic has seen quite extensive coverage in academic papers but getting it to work in the limitted environment of Flash provides some intersting challenges. Interesting enough for me to give it a try over my christmas break.
I loosely followed a paper Particle-based Viscoelastic Fluid Simulation to implement a sandbox application that simulates and renders a small blob of fluid and allows you to tinker with various parameters. Click the image below and give it a try!
The fluid is represented as a set of ~400 particles that interact with their neighbours and such seem to form a coherent entity.
For rendering I use an oldschool approach: Each particle is represented as a metaball. For a given pixel the force function of every metaball is evaluated, summed up and the final sum is compared with a threshold value to decide whether the pixel is inside or outside of the fluid.
To be able to do that in realtime I precalculate the values that a forcefunction yields for pixels within a relevant range of the particles center as a Bitmap.
These tiny images (about 40×40 pixels) are rendered at the position of the particle with additive blendmode.
Now I can decide with a single lookup in the resulting Bitmap if a pixel belongs to the surface or not. That evaluation step is done by a pixel bender shader and instead of just working with a boolean state (inside or outside) I sample a gradient based on the accumulated force value.
This approach allows me to achieve a variety of styles in the visual appearance of the simulated fluid by just using different gradients.
This technique would unfold it’s real potential if you could use hardware acceleration to accumulate the field functions because that’s a bottle neck in the current implementation.
Not sure if that stuff has any practical relevance, anyway. But the SWF is only 32kb big so people that are into demos (32k, 64k) could maybe make something fancy out of it. Here’s the source code released under the MIT license.
Developing and Selling Rune Hunt
I started experimenting with real time shadows in Flash in fall 2009. At the same time I took notice of the Flixel framework and all the cool retro games that people made with it. Why not make a game of my own that combines pixel graphics with dynamic light and shadows?
At the 14th December 2009 I created a thread in the Flixel forums with a mission statement:
The setup is classic: a hero in a dungeon full of treasure, traps and NPC of questionable attitude. The twist is that I’m limiting the player’s view to what his avatar is seeing. Top-down-ego-perspective so to speak. I hope it’s making exploring the dungeon more interesting and the enemies more scary if you don’t see all the stuff long before it becomes relevant/dangerous.
It’s been the beginning of an interesting journey…