Last weekend (6-8 September 2013) in Gdańsk University of Technology there was 3rd edition of WGK - Polish Conference on Computer Games Development. Just like in previous years, it was good and it's getting better and better. Many presentations were shown in up to 3 parallel sessions, so obviously everyone could see only a portion of them and my short review will be very subjective. You can also take a look at my photos from the conference.
There were two general kinds of presentations - scientific and industry. From the presentations made by game industry professionals, some were business oriented. That includes the discussion panel with Berlin delegation, as well as Saturday's "Lecture of the Day" with the outlook of European game market. Dates, percents, numbers, amounts, company names... - for me, that was quite boring, too much business and too little technology.
On the other hand, the presentation "Like a Boss" by Tomasz Gop and Michał Kuk from CI Games was great. They showed actual gameplay (played in real-time on a powerful PC) from their upcoming new titles - Alien Rage and Lords of the Fallen. The presentation was about designing bosses that players would like, so they presented boss fights and even showed some design docs of these bosses.
From the scientific lectures - the ones that also have a paper printed in the conference materials - I mostly remembered these two: Krzysztof Kluczek researched a way of animating skeleton of a 3D model using a set of control points connected with constraints on distance (something like springs) and also connected to an invisible triangle/tetrahedron to preserve general shape, all updated using Verlet method. He developed a tool where he can easily apply different kinds of presets and make a character of any skeleton topology run, attack or even dance with quite complex choreography (which looked very funny :)
Second interesting scientific lecture I've seen was the one presented by developers of Gizarma - a browser strategy game. They researched an algorithm for map generation. Based on sketched outline of island borders and rivers plus some density maps (like temperature, precipitation) and using some geography-based models, an iterative algorithm generated tesselation into irregular regions with desired properties and some parameters for each region (like the type of terrain). The execution of this algorithm can tak hours and we concluded that something similar could probably be achieved quickly (or even real-time) by starting with poisson disc, calculating Voronoi diagram and then applying some noise to its edges, but anyway their method is interesting and gives good-looking result.
Some papers were presented on the conference as posters. One of them caught my attention. Authors of an article "A role playing game name generator learning its creativity from Arkadia MUD players" developed an algorithm for generating new, yet reasonable names for fantasy RPG characters based on the concept of Markov Chains.
A very unusual presentation was showed by Maciej Miąsik. As a veteran of Polish game industry with 20+ years of experience, he talked about the value and the ways of preserving our old game projects and the history of their development, so we can go back to it when necessary - even many years later, when hardware platforms and storage devices change. He gave many arguments why it's worth doing, as well as gave some specific advices, e.g. to backup all the email history. I don't share his beliefs on the big value of the history, but anyway that was a very interesting lecture.
There was also a discussion panel about crunch time, involving several Polish game developers, moderated by Piotr Gnyp. They quite agreed that the crunch is bad, but general conclusion from their discussion could be that the crunch is always present, in every project. So they talked about how to predict, manage and compensate it, but not how to avoid it. Maciej Miąsik shared an interesting point of view - he said he studies this problem a lot and he currently believes that the crunch in gamedev is the result of making same mistakes that other businesses of software development made many years before. I asked a question about why do game developers - even those who strongly dislike crunch - usually stay until the end of the project and then leave, instead of leaving the job in the moment the manager announces start of the crunch. They said that developers generally "like to finish something", but I somehow feel that this answer is not fully exhaustive and satisfactory :)
There were also some workshops on the conference (e.g. about concept art or Unity), but I didn't attend any of them. Another part of the conference were integration parties. In both Friday and Saturday evenings, we were invited into quite elegant venues to have some beer and socialize. As always on such events, I met many friends from gamedev industry and Warsztat community, who came from all over the country. From our conversations, a general trend that remains in my mind is that many of them started their own business instead of having a stable, full-time job in some corporation. They are quite successful in terms of money, but they don't always claim to work less hours per week than they would in a traditional job. They also don't make their money from games (although they want to), but from different kinds of software projects. An outstanding exception is Daniel Sadowski with his company Nitreal Games. They started with doing games for government contracts, like Misja Bielany promoting Bielany - a Warsaw district, but now they made their own game (a time management casual PC game) - Gardens Inc. which is very successful.
Another part of the event was Developers Showcase - a place where amateur, as well as professional game developers could show their projects. I was amazed by the stands prepared for the exhibitors - they looked like on some kind of a small expo. The game that won voting on the best project was SuperHot - an FPS based on a novel idea that time passes (bullets fly, enemies walk) only when the players moves or rotates. The one project that was not a game used 4 Kinects to build a 3D model of objects so everyone who entered their space could see himself from any side as a colourful point cloud.
Finally, on Sunday there was the traditional game development competition (named Games Bonanza), where teams of up to 4 people had to develop a game in 8 hours. Unity and similar technologies were prohibited, so teams made their games in programming languages like C++ or Python with PyGame. The theme this year was "Spaghetti". I expected most games to be about the Flying Spaghetti Monster (especially after the lecture about designing boss fights), but the games turned out to be more diverse. One was about a cook avoiding spaghetti thrown at him by angry restaurant clients, another was about a piece spaghetti running away from the cook... At the end there were 5 entries in the competition. We took the 1st place :D Our team was: Krzysztof Kluczek "Krzysiek K." (programming, team leader), Tomasz Dąbrowski "Dab" (programming, level design), Michał Rudnicki "Mildanach" (graphics) and me (programming, audio). We made our game in Visual C++ 2010 Express based on a framework prepared by Krzysiek K., which uses DirectX 9. We created a 2D platform game where a plumber jumps, collects points and finally fights a boss. Here are some screenshots:
tl;dr: WGK is a nice conference for game developers with many different events taking place and it's definitely worth attending. By the way, in 2 weeks (20-22 September) there is different kind of party - WeCan demoparty in Łódź.