Rotational matrix as an example of Data Driven Design.

Every self respecting linear algebra teach causes students to hate matrices. If you can't you are not qualified. Only now, after three years without formal study of matrices, I learned to appreciate them.

I made a small computer graphics project at school. I had a separate function for X axis rotation, separate function for Y axis rotation, Z rotation and XYZ scaling were separate function, so was XYZ translation. My teacher told me to use matrices but I simply could not understand why it is better. I mean, matrices are creepy and functions are simple. Rotation around arbitrary axis was a bit cheaty (and buggy); this is fine for a small project though.

Now I understand the beauty of matrices. Now I see what is wrong with transformations stored as list of function objects. 8 years ago I was too young to understand the value of simple matrix that replaces arbitrarily long list of functions with parameters.

My friend recently showed me a piece of code from his hobby project. The code was full of conditional and switch cases. After little thought we realized that confusing branching in code can be replaced with a single matrix. It was the most satisfying change I ever made in code.

The value of well placed goto.

Some legacy programming languages do not support nor destructors nor exceptions. You must carefully check the returned error codes in case of error you must carefully free all resources. Often this leads to code like this:

error_code =some_func(some_params);
if (error_code!=NOERROR) {
  goto cleanup;
}
error_code =other_func(some_params);
if (error_code!=NOERROR) {
  goto cleanup;
}
error_code =yetanother_func(some_params);
if (error_code!=NOERROR) {
  goto cleanup;
}
//... More code here

There is much code like this. At some point a well-intentioned developer who heard that goto is evil changed the code to look like this:

do {
  error_code =some_func(some_params);
  if (error_code!=NOERROR) {
    break;
  }
  error_code =other_func(some_params);
  if (error_code!=NOERROR) {
    break;
  }
  error_code =yetanother_func(some_params);
  if (error_code!=NOERROR) {
    break;
  }
  //... More code here
} while (0);

In my opinion the do-break-while(0) is much uglier and harder to read than goto.
1. Here break is just a goto in disguise.
2. If you see the beginning but not the end you think there is a loop here.
3. In original version goto is used to control the flow in exceptional cases. Goto does not change the structure of the code under non-exceptional scenario.

What do you think?

More about globals.

In addition to possibility of an API that silently relies on global variables there are more potential pitfalls.

Suppose that you have a global variable which is modified within ten different if-else blocks. You wrote two tests - in one all ifs are true, in another all ifs are false. Now you believe that you have 100% code coverage.

The real coverage, however is not so good. 10 ifs that manipulate the same globals are semantically similar to a huge switch-case statement with 2^10=1024 possible cases. It will take 1024 tests to give true full coverage.

assert(g_everything->isEvil())

I'm in development for 3 months now. I must learn I understood something important. Something that I learned long ago but understood only now. Global vars are evil.

I saw many books about it, but I did not understand. A well placed global variable is very convenient and there is no harm in it. This is what I though when the largest project I experienced was around 2000 lines of code.

If you think what I thought 3 months ago you will hopefully learn from this blog post.

Look at this function signature:

error_code_t loadItemFromDefaultDir(const std::string filename);

As you can see function loads item and returns error code. Where is the item?

if(loadItemFromDefaultDir("it42.item")!=NO_ERR) {
  throw load_exception;
}
sub_item_t* mySubItem = g_everything->repository->collections->getSub("it42", MY_SUB_CODE);

You may think there is no problem here, and you are partially right. The problem begins when the item is loaded in one place but the function that uses it is in different module. And the loader does not check the error code. And the item name is not constant, but dynamically created string... You got the idea.

Sometimes you write code like this and it does not work for your argument, but works for any other argument.

convertAndNormilize(&myConvertibleString);

You spend much time looking at the code and finally veteran developer feels pity on you and explains that the string you pass is defined in g_everything->exceptions->conversion_exempt_strings and you must first call

g_everything->exceptions->override_convert_exc_when_norm(myConvertibleString)

C++11 factory

Today I was playing with awesome new features of C++11. In the examples below B is base class and D is derived.
The following code works fine:

std::function<std::shared_ptr<B>()> factory;
factory = [](){return std::make_shared<D>();};

The following code doesn't work:

auto factory = std::make_shared<B>;
factory = std::make_shared<D>;

The first code works because polymorphic wrappers for function objects are smart; the second does not work because function pointers are not smart at all. You can not cast from function that returns derived shared pointer to function that returns base shared pointer, you need function wrappers for it.
I tried to implement parametrized factory method. VC10 does not support initializer so the following did not work yet:

std::map<std::string, std::function<std::shared_ptr<B>()>> factory { 
    std::make_pair("D", [](){return std::make_shared<D>();}),
    std::make_pair("D1", [](){return std::make_shared<D1>();})
};

I'll test this trick with gcc4.6 tomorrow. In VC10 you can initialize the map by hand:

std::map<std::string, std::function<std::shared_ptr<B>()>> factory;
factory.emplace(std::make_pair("D", [](){return std::make_shared<D>();}));
factory.emplace(std::make_pair("D1", [](){return std::make_shared<D1>();}));

Please note the use of emplace method.
You can use factory like this:

auto Bptr = factory["D"]();
//Do something with D object...
Bptr = factory["D1"]();

Unfortunately std::make_unique is not a part of new standard, you can copy the code for make_unique from Sutter's blog. Factories returning unique_ptr are considered better.

Potential girlfriend match test.

After seeing so many GF tests I decided to make my own.

As she rides home from university alone in the bus observe her behavior. If she endlessly chats on cell phone it means she is ready to throw money away, she does not care about it, she is unable to count her own money. She is not serious and she is a bad match.

Limitations: Impossible if she does not use mass transit.

Disclaimer: it is a joke, do not take it seriously.

Is there any wealth created by clicking ads?

Recently many economics bloggers, some of them have Ph.D, blamed new financial instruments for the crisis. Key claim of many of them is that fancy financial instruments do not create zero real economic value are not connected to labor productivity. Something that gives one money without useful work done is potentially evil. I'm not going to argue with them, I will explain how intangible banner clicks generate wealth.

Billy and Jimmy create state-of-the-art decorative fences from beer cans. Give them soldering iron and a ton of recyclable aluminum and they will create real masterpieces. They are very proud dudes, so proud that they refuse to work with Coca-Cola cans, only beer cans, please, Czech bear preferred. Also they are not very successful at selling their product, so they only work 3 hours a day.

Alice browsed a web site and saw a Google ad (under Firefox with AdBlock extension she wouldn't see it). She clicked the banner to read about decorative fences. After reading the relevant information she decided to order some for her garden. For the next three week Billy and Jimmy work full time and overtime thus reducing unemployment and creating real product to be consumed.

This way a single click that increased employment, GDP, consumption and everything else.

But wait! Alice could spend the money differently, she could spend it on aesthetic wallpapers from synthetic waffles that would require even more work-hours without wasting precious aluminum. She could donate it to kids starving from cancer. She could save the money for later use.

One more thing to note. Bob and Rob also clicked the link, but they did not need anything that Jimmy and Billy could offer. Their clicks that cost real money to Jimmy without creating actual wealth, they are potentially evil.

Back to our topic. Was there any wealth created by clicks? We will never know. Economics is about making theories based on wrong assumptions. Assumptions that are known to be wrong from the beginning. You can define wealth in one way and discover five alternative ways and no definition is better, in general case, than the other.