Updated in November 2024 to include Crossroads of Ravens.

Without beating around the bush, the reading order for the Witcher books is as follows:

π is one of those constants which pops up when least expected. At the same time it’s sometimes missing when most needed. For example, consider the following application calculating area of a disk (not to be confused with area of a circle which is zero):

#include <math.h>
#include <stdio.h>
#include <stdlib.h>

int main(int argc, char **argv) {
	for (int i = 1; i < argc; ++i) {
		const double r = atof(argv[i]);
		printf("%f\n", M_PI * r * r);
	}
}

It uses features introduced in the 1999 edition of the C standard (often referred to as C99) so it might be good to inform the compiler of that fact with a -std=c99 flag. Unfortunately, doing so leads to an error:

$ gcc -std=c99 -o area area.c
area.c: In function ‘main’:
area.c:8:18: error: ‘M_PI’ undeclared (first use in this function)
    8 |   printf("%f\n", M_PI * r * r);
      |                  ^~~~

What’s going on? Shouldn’t math.h provide the definition of M_PI symbol? It’s what the specification claims after all. ‘glibc is broken’ some may even proclaim. In this article I’ll explain why the compiler conspire with the standard library to behave this way and why it’s the only valid thing it can do.

In a past life I’ve talked about a challenge to write the shortest program which prints all prime numbers less than a hundred. Back then I’ve discussed a 60-character long solution written in C. Since Rust is the future, inspired by a recent thread on Sieve of Eratosthenes I’ve decided to carry the task for Rust as well.

To avoid spoiling the solution, I’m padding this article with a bit of unrelated content. To jump straight to the code, skip the next block of paragraphs. Otherwise, here’s a joke for ya:

Continuing the new tradition of clickbaity titles, let’s talk about explicitness. It’s a subject that comes up when bike-shedding language and API designs. Pointing out that a construct or a function exhibits implicit behaviour is often taunted as an ultimate winning argument against it.

There are two problems with such line of reasoning. First of all, people claim to care about feature being explicit but came to accept a lot of implicit behaviour without batting an eye. Second of all, no one actually agrees what the terms mean.

In this article I’ll demonstrate those two issues and show that ‘explicit over implicit’ is the wrong value to uphold. It’s merely a proxy for a much more useful goal interfaces should strive for. By the end I’ll demonstrate what we should look at instead.

Updated in October 2021 to include direct comparison shift usage between Dvorak and Programmer Dvorak layouts.

A few years age I’ve made a decision that had the potential to change the course of history. Had I went a different path, the pl(dvp) layout might have never seen the light of day. But did I make a wise choice? Or had I chosen poorly?

I’m talking of course about the decision to learn Programmer Dvorak rather than a regular Dvorak keyboard layout. The main differences between the two is that in the former digits are entered with Shift key pressed down which allows several punctuation marks often used when programming to be typed without the need to reach for Shift. The hypothesis goes that developers use digits less often thus such design optimises the layout for them.

To test this I’ve grabbed all my git repositories and constructed a histogram of characters used in text files present there. Since letters are on the same position on both layouts in question, only digits and punctuation characters are compared on the histogram:

Some years ago, during a friendly discussion about C++, a colleague challenged me with a question: what’s the best way to represent a sequence of numbers if delete operation is one that needs to be supported. I argued in favour of a linked list suggesting that with sufficiently large number of elements, it would be much preferred.

In a twist of fate, I’ve been recently discussing an algorithm which reminded my of that conversation. Except this time I was the one arguing against a node-based data structure. Rather than ending things at a conversation, I’ve decided to benchmark a few solutions to make sure which approach is the best.

The problem

The task at hand is simple. Design a data structure which stores a set of words, all of the same length, and offers lookup operation which returns all words matching globs in the form ‘prefix*suffix’. That is, words which start with a given prefix and end with a given suffix. Either part of the pattern may be empty and their concatenation is never longer than length of the words in the collection. Initialisation time and memory footprint are not a concern. Complexity of returning a result can be assumed to be constant.

In this article I’me going to describe possible solutions — some using a boring vector while others taking advantage of an exciting prefix tree — and benchmark the implementations in an ultimate battle between contiguous-memory-based and a node-based containers.

‘I’m using slock as my screen locker,’ a wise man once said. He had a beard so surely he was wise.

‘Oh?’ his colleague raised a brow intrigued. ‘Did they fix the PAM bug?’ he prodded inquisitively. Nothing but a confused stare came in reply. ‘slock crashes on systems using PAM,’ he offered an explanation and to demonstrate, he approached a nearby machine and pressed the Return key.

Screens, blanked by a locker a few minutes prior, came back to life, unlocked without the need to enter the password.

I’ve written about L*a*b* so it’s only fair that I’ll also describe its twin sister: the L*u*v* colour space (a.k.a. CIELUV). The two share a lot in common. For example, they use the same luminance value, base their chromaticity on opponent process theory and each of them has a corresponding cylindrical LCh coordinate system. Yet, despite those similarities — or perhaps because of them — the CIELUV colour space is often overlooked.

Even though L*a*b* is getting all the limelight, L*u*v* model has its advantages. Before we start comparing the two colour spaces, let’s first go through the conversion formulæ.