
Welcome to the
Forge
There are three types of algorithms in the world.
The first type works. It does what it is told to do, without question and without surprises. This type is rare. You could say it is as rare as a taxi driver who doesn't talk.
The second type usually works. It does what you tell it to do – except on full moon nights, public holidays, and whenever the board is watching. This type is common. Very common, in fact. You don't even want to think about how common.
The third type doesn't work at all, but it was written by someone who got promoted, so it can't be touched. We remain silent about this type out of politeness.
We are responsible for the second type. (For the third type too, by the way, but that costs extra.)
At Forgotten Forge, we optimize algorithms. We make slow things fast and unreliable things reliable. That sounds modest, and it is. But it is a modesty that hides a certain competence – just as the modesty of a good cook usually hides a very sharp knife.
We don't promise you a revolution. Revolutions are loud, messy, and usually poorly organized. We promise you something much better: results. Quiet, clean, measurable results. The kind of results that make your boss think he underestimated you. (He probably did, but that's another story.)
About algorithms, butlers, and the nature of order
You might think that an algorithm is nothing more than a series of instructions that a computer processes. That's true. Just as you might think that marriage is nothing more than a contract between two people. Technically correct, but still missing the point.
A good algorithm is like a good butler. You don't notice it. It does its job quietly, thoroughly, and without fuss. It brings tea at the right time, puts the newspaper in the right place, and occasionally saves the honor of the house without ever saying a word about it. Only when it is missing—or when it is replaced by a bad butler—do you realize what you had in it.
Most algorithms we encounter are not butlers. They are more like that one intern every office knows: enthusiastic but unpredictable. Sometimes brilliant. Sometimes disastrous. Mostly somewhere in between, in a way that keeps you awake at night.
We turn interns into butlers. Admittedly, that's an unusual job description, but it gets to the heart of the matter. The difference between a mediocre algorithm and a good one rarely lies in the big picture. It lies in the small details. In the loop that takes three milliseconds too long. In the data structure that works when the weather is fine but collapses like a cheap umbrella when the wind picks up. In the query that someone wrote four years ago, which no one knows what it does anymore, and yet no one deletes it – out of superstition, probably.
It's always the little things. In code as in life. The devil is in the details, and he has made himself comfortable there.

Get Your Algorithm Forged
Some code runs. Some code performs.
And some code is forgedd.
There is code that runs. There is code that accomplishes something. And there is code that has been forged.
The difference? Forged code has been through something. It has been taken apart, examined, criticized, improved, and put back together again—with more care than most people invest in their tax returns. (Which, when you think about it, isn't a particularly high standard. But you get the point.)
We don't take on every problem. That's not arrogance—it's honesty. Some problems need a forge. And some problems need a long walk and maybe a chat with a friend. We're happy to help you figure out which kind you have. But you'll have to have that chat with your friend yourself. The process is very simple, because we have found that complicated processes usually serve to distract from the absence of actual work: You describe your problem to us. We take a look at it. If it's suitable for the forge, it gets forged. And what comes back doesn't just work. It works as if it had never intended to do anything else. Some people call this optimization. We call it decency. Because an algorithm that runs poorly, even though it could run better, is basically a small deception. And no one here likes deception, even small ones.
Write to us
If you've read this far, you're either interested or very bored. In either case, feel free to write to us. Those who are interested will receive a reply. Those who are bored will too – because politeness costs nothing, and we have plenty of it.
Note: The humor in these texts is intentional. So is the expertise behind them.


