What Makes a Craftsman?

Not the titles, not the fancy tools. It is not the expertize, either. It is the responsibility one takes in doing one’s job. Why? Because what one wants from a craftsman is to have one’s problem solved. You want a doctor to see you through the illness and make you feel better, a developer -- to write software that serves your purpose, etc. You want the result.
There always was and always will be a tendency to consider the tools and titles when faced with a choice, which implies their necessity to you as a craftsman. If you're a developer, you'd better have all those fancy certificates saying you really know the languages, platforms, methodologies etc. People expect it. It will help you get your customers. Yet if you don't care, if you take no responsibility or don't provide the result they want, no certificate will help you keep them.
That is exactly why the most important part of your CV is the one saying what have you done. That is why, when choosing a doctor one checks his customer reviews rather then his diplomas. That is the ultimate quality criterion.

The job I'm sabotaging at the moment involves machine learning for probabilistic models. Such are necessary when the problem cannot be solved deterministically either due to high computational complexity or due to lack of our knowledge of the actual structure of the problem. One example of the latter case is the natural language, which is quite complex and inherently ambiguous. Linguists, psycholinguists and cognitive scientists keep generating and checking numerous hypotheses on the subject, and yet there's no single comprehensive model of the language, no algorithm to pass the Turing test. So we stick to small subtasks we can manage, and even there we generally design a simplistic model reflecting our idea of the underlying structure and leave a number of degrees of freedom in the model to be consequently adjusted to fit the data. The process of adjustment is called training. The idea is simple -- we provide a number of sample inputs and try to find parameter values which allow the model to best reproduce the correct outputs. The examples used for training are called the training set.
Sounds simple, yeah. But there's a number of catches, of course. Firstly, the performance of the model on the training set is only partially indicative of its expected performance on the other, unseen data. Besides, depending on the structure of the model, the training can be an extremely computationally intensive task -- training can usually be seen as an optimization process, not necessarily convex and usually in a high-dimensional space. Finally, the more complex the problem -- the bigger training set we need. And training data may be expensive, since it often takes a skilled linguist or other specialist to produce it.
Funny, really. Most humans are exposed to only a limited number of training examples and still manage to learn well from them. We still don't exactly know how.

Language != Location

Dear Web Designers and Usability Guys,

More and more sites are available in multiple languages these days. And some even provide different information based on the location of the user, e.g. relevant news, local weather, or price information.
For the majority of Internet users, location implies language. However, it is not always the case! If you are only interested in language, ask the user which one does he or she prefer. If you are interested in location - ask for location. It is cool if you are able to detect my location automatically, but it drives me mad if I can't change it manually if needed. Suppose I want to look up the news for a different place? Or I'm planning to buy this device when I'm back home? Why can't I look up the bloody price I'll get there?
And remember: location doesn't imply language! For example, I currently stay in Germany, but I don't really speak the language, and it's really annoying when I cannot change it. A classic example is Amazon: when I'm in Germany, I'd like to search amazon.de and see the products they have here and the local prices in Euros. I can do that, but only in German. I cannot switch to English without switching to GB or USA.

Cognitive Science

Cognitive Science (CogS) is an extremely fascinating and rapidly developing area of human knowledge. CogS is to answer questions about the thing people are interested in most of all - about themselves. The questions are: how do we perceive, process and transmit information? And how do we learn?
It may be considered a specific part of psychology. A rather independent part, though. One special aspect is it's close relation to computer science and machine learning. Not just to understand, but also to imitate the processes similar to human thinking in computer algorithms - there are still many things computers cannot do, while for humans they are so obvoius we don't even think of them. On the other hand, if we construct an algorithm that performs close to human on certain tasks, it can yield a good model for what actually happens.
I was lucky to participate in MLSS 2010 last week, where I got an introduction to the field via several amazing talks by Josh Tenenbaum, Tom Griffiths, Chris Watkins and others.
Personally, I believe that knowing how the brain works, we could not only construct better computer systems, but also improve the way we learn and teach.